The Role of Thermochemical Processing in Future Biorefineries

Final Report

Iowa Energy Center

Iowa Energy Center Grant No.: 0606

Report Period: July 1, 2006 – June 30, 2008

Principal Investigator: Robert C. Brown

Organization: Iowa State University

Signature of PI

The Role of Thermochemical Processing in Future Biorefineries

Principal Investigator: Robert C. Brown

Executive Summary

The overall goal of this project was to evaluate the role of thermochemical processing in the emergence of advanced biorefineries. This goal has been accomplished through the preparation of three papers on the technical and economic aspects of advanced biorefineries. An extensive review of biorefinery technologies has been completed and is included in this report. Two journal articles, “Comparative Economics of Biorefineries Based on the Biochemical and Thermochemical Platforms” and “Establishing the optimal sizes of different kinds of biorefineries,” have been published in Biofuels, Bioproducts, and Biorefining, These studies show that thermochemical conversion, as well as hybrid biological/thermochemical platforms, holds great potential in expanding biomass conversion pathways. The review paper “Routes to Biofuels” describes the variety of pathways for conversion of biomass to transportation fuels. Biological, thermochemical, and hybrid biological/thermo-chemical platforms are all included in this report which takes into consideration potential biofuel applications and limitations. This report also draws from previous analyses to determine the economics of these technologies. This project reviewed the current biorefinery concepts and identified potential roles for thermochemical processing; evaluated the current status of biomass-derived transportation fuels, and assessed the technical feasibility of various methods proposed for the conversion of biomass to fuels. “Comparative Economics of Biorefineries Based on the Biochemical and Thermochemical Platforms” describes a technoeconomic study that compares biochemical and thermochemical processing. This comparison took into consideration technology maturity, plant capacity, energy content of the fuel, feedstock costs, method of calculating capital charges, and year in which the original analysis was assumed. Results from this analysis showed that capital costs for advanced biochemical and thermochemical biorefineries will require capital investments of four to five the capital of comparably sized grain ethanol plants. Nevertheless, fuel costs from advanced biorefineries will be similar to grain ethanol at feedstock prices of $3.00 per bushel of grain. A second technoeconomic study, “Establishing the optimal sizes of different kinds of biorefineries” explored the factors that influence the optimal plant size and resulting fuel cost of advanced biorefineries. Common economies-of-scale calculations predict lower unit production prices for large facilities. Biomass facilities producing hundreds of millions of gallons of fuel will require transportation of tens of thousands of tons of feedstock daily generating a significant delivery expense. The combination of these factors establishes an optimal plant size at which fuel production costs per gallon of gasoline equivalent reaches a minimum value. This study concluded that advanced biorefineries converting lignocellulosic biomass would be optimally sized at 240 to 486 million gallons of gasoline equivalent compared to 79 million for grain ethanol plants.

Routes to Biofuels Robert C. Brown and Mark Wright

1 Introduction

The current generation of biofuels is based on grain ethanol and biodiesel from soybeans or other high-oil content plants. However, it is unlikely that corn and oil crops can provide more than a small fraction of world-wide demand for transportation fuels, displacing no more than about 12 billion gallons of gasoline in the United States, for example, whereas the national target is 60 billion gallons of ethanol by 20301. Cellulosic (fibrous) biomass will likely be required to make up the difference.

Much of the discussion about advanced biofuels centers around cellulosic ethanol, assumed to be produced by enzymatically hydrolyzing hemicellulose and cellulose into five-carbon and six carbon sugars that can be fermented into ethanol.2 In fact, a number of advanced biofuels are possible including hydrogen, butanol, methanol, mixed alcohols, esters, ethers, Fischer-Tropsch liquids, and so-called green diesel produced by hydrocracking vegetable oils, bio-oil, and biocrude. Accordingly, it may be too early to select the “advantaged molecule(s)” for use as advanced biofuels or the processes by which to produce these future fuels.

This paper reviews the various fuels under consideration for production from biomass and the biochemical, thermochemical, and hybrid processes that could produce them. In instances where relatively complete techno-economic data is available the various processes are compared in terms of optimal size, capital costs, and operating costs.

2 Potential Biofuels

Biobased transportation fuels, also known as biofuels, are currently dominated by ethanol and mono alkyl esters (biodiesel). However, there are other candidate liquid biofuels including methanol, butanol, mixed alcohols, Fisher-Tropsch liquids, methylated furans, and so-called green diesel, as well as gaseous biofuels including hydrogen, methane, ammonia, and dimethyl ether. Ethanol is currently produced by fermentation of glucose derived from starch crops or sugar crops. Fermentation can also produce hydrogen, methane, butanol, and even mono alkyl esters, although none are widely produced biologically for use as transportation fuels at present.

Alternatively, thermochemical processes can be used to produce biofuels. The major thermochemical pathways include gasification, fast pyrolysis, hydrothermal processing, and catalytic dehydration of carbohydrates. The first of these produces a gaseous product, fast pyrolysis yields mostly a liquid product, hydrothermal processing can produce either gaseous or liquid products depending upon the processing conditions, and catalytic dehydration produces liquids. Both gaseous and liquid products typically require upgrading before they are suitable as transportation fuel. Upgrading processes include syngas catalysis, syngas fermentation,

1

2

hydrocracking of vegetable oils and thermally-derived liquids, and catalytic dehydration of carbohydrates to furans.

Biomass gasification is the process in which plant materials are thermally decomposed into a gaseous mixture consisting primarily of carbon monoxide (CO) and hydrogen (H2) known as producer gas if it contains substantial quantities of nitrogen or syngas if it contains little nitrogen.3 Nitrogen content depends upon whether the solid fuel feedstock is processed in air, oxygen, or in the complete absence of oxygen. Producer gas or syngas can be reacted with steam to produce hydrogen fuel or passed over metal catalysts to synthesize ethanol, mixtures of alcohols, Fisher-Tropsch (F-T) liquids, or dimethyl ether. This gas can also be fermented by certain bacteria into hydrogen, ethanol, and esters.4 Gasification at high pressures and low temperatures can also be used to generate methane (CH4). Usually heat and pressure are the driving forces of gasification but sometimes catalysts are employed to allow processing at lower temperatures and to improve product selectivity.5

Fast pyrolysis is the rapid heating of biomass in the absence of oxygen to generate a liquid product known as bio-oil along with smaller amounts of gas and char. The bio-oil consists of a carbohydrate-derived phase and a lignin-rich phase.6 Bio-oil can be used directly as boiler fuel or burned in appropriately modified stationary diesel engines or gas turbines, but for transportation applications, it must be upgraded to a higher quality liquid fuel. This can be accomplished by gasifying the bio-oil to syngas followed by catalytic transformation to Fisher-Tropsch diesel. Alternatively, upgrading can be achieved by hydrocracking the bio-oil to hydrocarbons7 that are sometimes known as green diesel to reflect its origin.

Hydrothermal processing employs a high pressure, aqueous-phase environment to convert biomass into gaseous or liquid products.8 The nature of the products depends upon the reaction conditions. The mildest processing conditions (200-250 °C and 50-100 bar) yields carbohydrate that can be hydrolyzed to fermentable sugars.9, 10 More severe processing conditions(300-350 °C and 120-180 bar) yields so-called biocrude, which contains a variety of organic compounds suitable for upgrading to gasoline or diesel fuel.11 The highest temperatures and pressures (600 °C and 230 bar) yield a gaseous mixture of hydrogen, carbon monoxide, and methane.12

Table 1 summarizes the properties of various biofuel options and compares them to fossil-fuel derived gasoline and diesel. The properties include specify gravity, kinematic viscosity, boiling point range, flash point, autoignition temperature, octane number, cetane number, heat of vaporization, and lower heating value. In the case of fuels that are gaseous at ambient conditions, specific gravity is reported for saturated liquids at their boiling point.

Tab

le 1

. C

ompa

riso

n of

tran

spor

tatio

n fu

els (

adap

ted

from

Ref

eren

ce 1

3)

*

Mea

sure

d at

16

C e

xcep

t for

liqu

efie

d ga

ses,

whi

ch a

re sa

tura

ted

liqui

ds a

t the

ir re

spec

tive

boili

ng p

oint

s.

Fuel

type

Fo

ssil

fuel

-der

ived

B

iom

ass-

deriv

ed

Gas

olin

eN

o. 2

D

iese

lFu

el

Met

hyl

este

r(f

rom

so

ybea

noi

l)

Met

hano

lEt

hano

lB

utan

olFi

sche

r-Tr

opsc

h A

2,5-

Dim

ethy

l fu

ran*

*H

ydro

gen

Met

hane

Dim

ethy

l et

her

Spec

ific

grav

ity *

0.

72-

0.78

0.85

0.88

60.

796

0.79

40.

810.

770

0.90

0.07

1(li

q)

0.42

2(li

q)

0.66

0(li

q)

Kin

emat

ic

visc

osity

at 2

0-25

C (m

m2 /s

) 0.

82.

53.

90.

751.

513.

642.

0810

516

.50.

227

Boi

ling

poin

t ra

nge

(C

)30

-225

210-

235

339

6578

117

164-

352

92-9

4-2

53-1

62-2

4.9

Flas

h po

int (

C)

-43

5218

811

1337

58.5

6.7

-184

-A

utoi

gniti

onte

mpe

ratu

re (

C)

370

254

-46

442

334

3-

NA

566-

582

540

235

Oct

ane

no.

(res

earc

h)

91-1

00-

-10

910

996

- 10

5 -

120

>130

>120

-

Oct

ane

no.

(mot

or)

82-9

2-

-89

9078

- 89

-

--

-

Cet

ane

no.

<15

37-5

655

<15

<15

2574

.6-

->5

5H

eat o

f va

poriz

atio

n(k

J/kg

)38

037

5-

1185

920

430

-44

750

940

2

Low

er h

eatin

g va

lue

(MJ/

kg)

43.5

4537

20.1

2736

43.9

4212

049

.528

.88

** P

rope

rties

repo

rted

in Y

Rom

án-L

eshk

ov e

t al,

Nat

ure

(200

7) 4

47, 9

82 e

xcep

t oct

ane

num

ber f

rom

Jam

es D

umes

ik, N

atur

ePo

dcas

t 21

June

200

7, h

ttp://

ww

w.n

atur

e.co

m/n

atur

e/po

dcas

t/v44

7/n7

147/

natu

re-2

007-

06-2

1.ht

ml.

3

2.1 Biodiesel

Vegetable oils, which are triglycerides of fatty acids, have long been recognized as potential fuels in diesel engines. Compared to petroleum-based diesel fuels, vegetable oils have higher viscosity and lower volatility, which results in fouling of engine valves and less favorable combustion performance, especially in direct-injection engines.14 The solution to this problem is to convert the triglycerides into methyl esters or ethyl esters of the fatty acids, known as biodiesel, and the byproduct 1,2,3-propanetriol (glycerol).

Table 1 illustrates that fuel properties of biodiesel are similar to petroleum-based diesel although biodiesel has a lower cloud point, which can cause fuel filters to plug under cold weather conditions. In general, oil crops from traditional agriculture are lower yielding than starch crops and biodiesel is more expensive than ethanol on an energy basis. For these reasons, vegetable oils are unlikely to make significant contributions to world-wide demand for transportation fuels unless dramatical improvements in crop productivity and process costs are achieved.

Oil palm grown in the tropics can yield over ten times as much vegetable oil as soybeans grown in temperate climates although there are concerns that rain forests are being cleared to expand oil palm plantations.15 Algal biomass is gaining increasing attention as a source of triglycerides because theoretical productivity in desert (high insolation) regions of the world is one hundred to two hundred times higher than for soybeans.16

2.2 Green Diesel

Diesel derived from renewable biomass is commonly known as green diesel. These are typically produced via gasification to syngas and catalytic upgrading to liquid fuels. Fischer-Tropsch synthesis is one of the most promising pathways for large scale production of green diesel, which comprises up to 40% of Fischer-Tropsch liquid products. Fischer-Tropsch diesel properties contribute to a clean combustion performance. Nevertheless, low density and low aromatic content result in lower energy content than petroleum diesel and unattractive blending properties.

Some of the concerns with green diesel vehicle use include complications with fuel injection systems, power output, and lubrication properties. Biomass diesel is capable of penetrating elastomers present in fuel injection systems due to the polarity of the diesel and base polymeric units of the elastomer17. Fischer-Tropsch diesel naphtenes and paraffins have a lower density and volumetric heating value than aromatics18, and this can reduce diesel engine power output19.Hydroprocessed diesel, both renewable and fossil-based, have poor lubrication properties caused by the reduction of surface-active polar compounds20, but this can be improved with additives.

A proposed solution to improve green diesel properties for use as a conventional diesel alternative is to blend biomass based diesel with coal based diesel.21 Although green diesel has high compatibility value with existing infrastructure and transportation system, it currently cannot compete economically with low sulfur diesel and may require large cost reductions.22

4

2.3 Ethanol

The biological production of ethanol and its distillation has been practiced for thousands of years. This fact combined with the wide availability of suitable fermentation feedstocks (grains and sugar cane) explains the early entry of ethanol into the biofuels market. In fact, Henry Ford had intended to fuel the Model T with ethanol, but inexpensive gasoline soon replaced it.

Ethanol has some shortcomings compared to gasoline. On a volumetric basis, it only has 66% of the heating value of gasoline. However, fuel economy depends on many complex interactions between a fuel and the combustion environment within an engine, which some argue improves the relative performance of ethanol23. For example, the higher octane number for ethanol compared to gasoline (109 vs. 91- 101) allows engines to be designed to run at higher compression ratios, which improves both power and fuel economy. Estimates for efficiency improvements in engines optimized for ethanol instead of gasoline range are 15% to 20%, resulting in a driving range approaching 80% of that of gasoline.24

A significant problem with ethanol-gasoline blends is water-induced phase separation. Water contaminating a storage tank or pipeline is readily absorbed by ethanol, resulting in a lower water-rich layer and an upper hydrocarbon-rich layer, which interferes with proper engine operation. Water contamination is a problem that has not been fully addressed by the refining, blending, and distribution industries; thus transportation of ethanol-gasoline blends in pipelines is not permitted in the United States and long-term storage is to be avoided.25

2.4 Methanol

Methanol is one of the ten most produced chemicals in the world. Methanol is a clear, odorless, flammable liquidthat can be generated from any hydrocarbon. Methanol production is a mature process with process efficiencies of 99% or higher. According to Davenport26, almost all methanol synthesis derives from natural gas because of cost considerations. It is an intermediate chemical in the production of formaldehyde, dimethyl ether, acetic acid, and others. Methanol can also be blended with gasoline for use as a clean transportation fuel.

As a transportation fuel, it has many of the same advantages and disadvantages as ethanol.27 The fuel properties of methanol are similar to those of ethanol: narrow boiling point range, high heat of vaporization, and high octane number. It has only 49% of the volumetric heating value of gasoline. However, methanol is considerably more toxic than ethanol. Recent rulings by the U.S. Environmental Protection Agency (EPA) are likely to ban the closely related and similarly toxic MTBE as a fuel additive because of concerns about ground water contamination.

A promising application for methanol is in fuel cell technology. Methanol is a leading candidate to provide the hydrogen necessary to power fuel cell vehicles, the next generation of automotive technology.28

5

2.5 Biobutanol

Before the rise of the petrochemical industry, butanol was traditionally produced by the acetone-butanol-ethanol (ABE) fermentation using Clostridium acetobutylicum. In a relatively complicated and difficult to control process, the ABE fermentation begins by producing butyric, propionic, lactic, and acetic acids, followed by a shift to butanol, acetone, isopropanol, and ethanol. The butanol yield from glucose is low, in the range of 15-25%. Butanol concentrations are typically lower than 13 g/L due to product inhibition. Because of these low yields and process challenges, butanol has remained an expensive fuel product.

Butanol has physical properties that make it more attractive than other alcohols for use as a motor fuel.29 Its volumetric heating value is 89% of gasoline, improving its fuel economy compared to ethanol and methanol. Its relatively low vapor pressure compared to other alcohols reduces problems with vapor lock and the emission of volatile organic compounds (VOCs) from fuel tanks. Although its octane rating is lower than ethanol and methanol, it is higher than gasoline. Butanol-gasoline blends are less susceptible to phase separation in the presence of water contamination, which should improve engine performance compared to ethanol and allow butanol to be transported through existing gasoline pipelines.In an attempt to revive the production of butanol from biomass, cell recycle, cell immobilization and extractive fermentation have been explored. Advanced fermentation technologies30 produce only hydrogen, butyric acid, and butanol, and have the potential to produce almost as many gallons of fuel from corn as ethanol fermentation 31,32

2.6 Mixed Alcohols

Efforts in Germany during World War II to develop alternative motor fuels discovered that iron-based catalysts could yield appreciable quantities of water-soluble alcohols from syngas33.These early efforts yielded liquids containing as much as 45-60% alcohols of which 60-70% was ethanol with lesser amounts of methanol, butanol, and other alcohols.

Some researchers have advocated the use of “mixed alcohols” as transportation fuels because the product typically contains a mixture methanol, ethanol, 1-propanol, and 2-propanol. One advantage is the ability to use lower H2:CO ratios than is required for methanol or Fischer-Tropsch synthesis.34 An extensive review of mixed alcohol synthesis technology is found in Reference 35.

2.7 Fischer-Tropsch Liquids

The technology was extensively developed and commercialized in Germany during World War II when it was denied access to petroleum-rich regions of the world. Likewise South Africa, faced with an world oil embargo during their era of apartheid, employed F-T technology to sustain its national economy. A comprehensive bibliography of F-T literature can be found on the Web.36

6

Fischer-Tropsch liquids composition depends on the process selectivity. To obtain high amounts of heavy hydrocarbons, a high liquid selectivity is required. Process selectivity is affected by various factors including catalyst and feed gas properties. The Anderson-Schulz-Flory (ASF) distribution describes the probability of hydrocarbon chain growth where the molar yield for a carbon chain can be calculated using the following equation:37

Cn = n-1(1- ) (1)

where is the chain growth probability of a hydrocarbon of length n. Light hydrocarbons (mostly methane) can be fed into a gas turbine to provide power. Fischer-Tropsch liquids can be separated into various products in a process similar to petroleum distillation. Hydrogen can also be used in hydrotreating, hydrocracking, and hydroisomerization processes to upgrade fractions of the Fischer-Tropsch liquids into common transportation fuels although there is a limit.

Senden et. al. calculated a theoretical limit for gasoline yield selectivity to be 48%38. Fischer-Tropsch diesel’s selectivity limit is estimated to be even lower at 40%. Remaining fractions can be used as a substitute to petroleum-based chemicals. Low selectivity towards gasoline and diesel makes it critical for Fischer-Tropsch derived chemicals to compete economically in the market.

2.8 Hydrogen

Hydrogen is often touted as an ideal fuel because it is clean burning and the combustion products include no greenhouse gases. It may be one of the most cost-effective ways to produce transportation fuels from biomass because a relatively small number of unit operations are involved in its production. Commercial production of hydrogen is dominated by steam reforming of natural gas although it can be produced from any carbonaceous fuel including coal and biomass:

CxHyOz + (2x-z)H2O xCO2 + (2x + y/2 –z)H2 (2)

This endothermic reaction either needs an external source of heat to drive the reaction, or oxygen needs to be added in sufficient quantities to support the exothermic reaction:

C + 1/2O2 CO (3)

in a process known as autothermal reforming.

Although this might be one of the most cost-effective ways to produce clean-burning biofuel, the physical characteristics of hydrogen present challenges in its use as transportation fuel. In particular, its low density even under cryogenic or high pressure conditions limits on-board storage of this fuel. Its wide flammability range also presents unique safety problems in its use in transportation systems39.

7

2.9 Methane

Methane (CH4) is a clean burning gas that is the main component of natural gas, widely used for residential and industrial heating applications and in electric power generation. Compressed natural gas is used in some urban mass transit applications40 and efforts are underway in Europe to use biomass-derived CH4 as renewable transportation fuel41. Liquefied natural gas (LNG) is sometimes produced to ease shipping of stranded natural gas to distant markets. However, most automotive applications contemplate its use as compressed gas.

Renewable methane from biomass can be produce by either biochemical or thermochemical processes. Biochemical production of methane (anaerobic digestion) is most suitable for high moisture, low value feedstocks such as manure, food processing byproducts, or waste water.Thermochemical production of CH4 (gasification) requires relatively low moisture feedstocks to avoid high thermal energy penalties.

Anaerobic digestion is the decomposition of organic wastes, including polysaccharides, proteins, and lipids, to gaseous fuel by bacteria in an oxygen-free environment42. The process occurs in stages, each involving specific types of bacteria, to successively break down the organic matter into simpler organic compounds. The desired product, known as biogas, is a mixture of about 60 vol-% CH4 and 40 vol-% CO2. It also contains traces of other gases, most importantly hydrogen sulfide (H2S), which must be scrubbed from the gas before it is used as fuel to prevent emissions of sulfur dioxide into the atmosphere.

Gasification can also yield methane by a process known as hydrogasification:43

C + 2H2 CH4 (4)

CO + 3H2 CH4 + H2O (5)

These exothermic reactions require low temperatures, high pressures, and large quantities of hydrogen. Separate generation of hydrogen and catalysts to achieve reasonable reaction rates are usually required to achieve high methane yields. It has been demonstrated at the commercial scale using coal as the carbonaceous fuel.

2.10 Dimethyl ether

Dimethyl ether, like liquefied petroleum gas (LPG), is a non-toxic, flammable gas at ambient conditions that is easily stored as liquid under modest pressures44. It is currently used as an aerosol propellant in the cosmetic industry, but has excellent potential as a fuel for heating, cooking, and power. It is particularly attractive as a substitute for petroleum-based diesel fuel since it has comparable cetane number but yields essentially zero particulate emissions and low NOx emissions. It is produced either catalytically from syngas

3H2 + CO CH3OCH3 + CO2 (6)

8

or through the dehydration of methanol by reactions at high pressure over catalysts.45,46

2CH3OH CH3OCH3 + H2O (7)

The direct reaction from syngas is thermodynamically more efficient. Either of these routes ultimately involves the production of syngas from carbonaceous feedstocks, including fossil fuels, and much of the current interest in this alternative fuel arises from the possibility of manufacturing it from inexpensive stranded natural gas.

2.11 Methylated furans

Furans are heterocyclic aromatic ethers consisting of a ring of four carbon atoms and one oxygen atom. Furans are colorless, water-insoluble, flammable liquids with volatility comparable to hydrocarbons of similar molecular weight. They were originally derived from the destructive distillation of wood although a number of synthesis routes are possible.47 Methylated furans have heating values and octane numbers comparable to gasoline making them attractive as transportation fuel.48 2,5 dimethyl furan in particular has received recent interest because new catalytic synthesis routes from sugars have been developed.49, 50 However, neither the fuel properties nor toxicity of this compound have been much studied raising questions as to the ultimate practicality as transportation fuel.

3 Biochemical Processing

3.1 Starch-to-ethanol

Starch is a polymer that accumulates as granules in many kinds of plant cells where they serve as a storage carbohydrate. Mechanical grinding readily liberates starch granules. The hydrogen bonds between the basic units of maltose in this polymer are easily penetrated by water, making depolymerization and solubilization relatively easy. Cereal grains also contain other components, such as protein, oil and fiber, which may be of sufficient value to recover along with the starch.

Hydrolysis, the process by which water splits a larger reactant molecule into two smaller product molecules, is readily accomplished for starch. Acid catalyzed hydrolysis in “starch cookers” at temperatures of 150 – 200 C proceeds to completion in seconds to minutes. In recent years enzymatic hydrolysis has supplanted acid hydrolysis due to higher selectivity.

The capital investment for dry milling is less than that for a comparably sized wet-milling plant. However, the higher value of its by-products, greater product flexibility, and simpler ethanol production can make a wet-milling plant a more profitable investment.

A typical dry milling plant will produce about 9.5-9.8 L (2.5-2.6 gal) of ethanol per bushel of corn processed. Yields of coproducts per bushel of corn are 7.7-8.2 kg (17-18 lb) of DDGS and 7.3-7.7 kg (16-17 lb) of carbon dioxide evolved from fermentation, the latter of which can be sold to the carbonated beverage industry. As a rule of thumb, the three products are produced in

9

approximately equal weight per bushel, with each accounting for approximately one-third of the initial weight of the corn.

3.2 Cellulose to Ethanol

Much of the carbohydrate in plant materials is structural polysaccharides, providing shape and strength to the plant. This structural material, known as lignocellulose, is a composite of cellulose fibers embedded in a cross-linked lignin-hemicellulose matrix 51. Depolymerization to basic plant components is difficult because lignocellulose is resistant to both chemical and biological attack 52.

Cellulose to ethanol consists of five steps: pretreatment, enzymatic hydrolysis, fermentation, and distillation.53 Of these, pretreatment is the most costly step, accounting for about 33% of the total processing costs54. An important goal of all pretreatments is to increase the surface area of lignocellulosic material, making the polysaccharides more susceptible to hydrolysis. Thus, comminution, or size reduction, is an integral part of all pretreatments. Some pretreatments are thought to reduce crystallinity of cellulose, which improves reactivity, but this does not appear to be the key for many successfully pretreatments.

Three basic methods for hydrolyzing structural polysaccharides in plant cell walls to fermentable sugars are available: concentrated acid hydrolysis, dilute acid hydrolysis, and enzymatic hydrolysis52, 55. The two acid processes hydrolyze both hemicellulose and cellulose with very little pretreatment beyond comminution of the lignocellulosic material to particles of about 1 mm size. The enzymatic process must be preceded by extensive pretreatment to separate the cellulose, hemicellulose, and lignin fractions.

Enzymatic hydrolysis was developed to better utilize both cellulose and hemicellulose from lignocellulosic materials. Pretreatment solubilizes hemicellulose under milder conditions than those required for acid hydrolysis of cellulose. Subsequent enzymatic hydrolysis of the cellulose does not degrade pentoses released during prehydrolysis. Cellulose is a homopolysaccharide of glucose linked by -1,4’-glycosidic bonds. Thus, enzymatic hydrolysis of cellulose proceeds in several steps to break glycosidic bonds by the action of a system of enzymes known as cellulase.The system of enzymes also usually contains hemicellulase to hydrolyze any hemicellulose not solubilized by prehydrolysis.

Simultaneous saccharification and fermentation (SSF) has been developed for fermenting sugars released from lignocellulose52, 55. The SSF process combines hydrolysis (saccharification) and fermentation to overcome end product inhibition that occurs during hydrolysis of cellobiose. By combining hydrolysis and fermentation in the same reactor, glucose is rapidly removed before it can inhibit further hydrolysis. The SSF process is illustrated in Figure 1. The biomass feedstock is milled and then prehydrolyzed to yield a mixture of pentoses, primarily xylose and arabinose, and fiber. The pentose is separated from the undigested fibers and fermented with appropriate yeast or bacterial to yield ethanol. The pretreated fibers are mixed with cellulases, which are either purchased commercially or produced on site, along with yeast and nutrients. The cellulose is solubilized to hexose (glucose), which is simultaneously fermented to ethanol. The rate-

10

limiting sused as bsugar or

step is the hyboiler fuel. Tstarch ferme

ydrolysis of The beer is dentations.

cellulose to distilled to et

glucose. Lithanol in a p

ignin is sepaprocess ident

arated from ttical to that e

the mixture aemployed aft

andfter

4 The

4.1 Ga

ThermolygasificatiGasificatmonoxidgas, can b

The syngmethanolproduce hare gaseotransportusually pvarious asulfur, chproduce t

The ratiosynthesisBiomass process.syngas an

Figure

ermoche

asificatio

ytic processiion (to prodution involves

de gaseous mbe further pr

gas gas can bl, ethanol, mhydrogen (Hous compountation fuels. preferred for amounts of lihlorine, and the desired p

o of H2 to COs favors a 2.1gasificationHydrogen e

nd steam ove

CO + H2O

1 Cellulose-

emical P

on

ing of carbonuce a gaseous the partial

mixture calledrocessed to n

be catalyticalmixed alcohoH2), methanends at ambieA relativelysynthesizing

ight hydrocaammonia, soproportions o

O in the synt15 to 1 ratio

n does not neenrichment, wer a catalytic

O CO2 + H

-to-ethanol

Process

n based matus product) ocombustiond synthetic gnumerous pr

lly processedls, and Fishe

e (CH4), diment conditiony pure mixtug these comparbons, tar, pome downstrof H2 and CO

thetic gas is while metha

ecessarily yiewhich increac bed, which

H2

processing

sing

erials can beor fast pyrolyof carbon-ri

gas (syngas).roducts.

d to a varietyer-Tropsch (ethyl ether (

ns but can beure of carbonpounds. Sincparticulate mream treatmO.

significant fanol can be peld the H2 toases the H2 th promotes th

(via enzyma

e generally cysis (to prodich materials. Synthetic g

y of liquid tr(F-T) liquids(CH3OCH3),e compressedn monoxide (ce raw produ

matter, and trent of the ga

for catalytic processed w

o CO ratio reto CO ratio, he water-gas

atic hydroly

categorized aduce primarils to a hydroggas, also kno

ransportations. It can also and ammond or liquefied(CO) and hyucer gas canace contamias stream ma

synthesis. Fiwith 3:1 H2 toequired for thcan be achies shift reactio

ysis)

as either ly liquid progen and carbown as produ

n fuels incluo be used to nia (NH3), wd for use as

ydrogen is n also containnants such a

ay be require

ischer-Tropso CO ratio35.he subsequeneved by passon:

11

oduct). bonucer

uding

which

nased to

sch.ntsing

(8)

Low temperatures thermodynamically favor this slightly exothermic reaction. To obtain satisfactory reaction rates, catalysts are employed in one or more fixed bed reactors operated in the temperature range of 250-400 C. It is possible to catalyze this reaction with a chromium promoted iron formulation56.

An alternative approach to obtaining neat ethanol from syngas is to first synthesize methanol and subsequently reacting this product with additional syngas:57

CH3OH + 2CO + H2 CH3CH2OH + CO2 (9)

Direct carbonylation of methanol has the advantage of yielding ethanol without coproduct water, which would eliminate energy-intensive distillations. The cost-effectiveness of this approach to ethanol synthesis has not been proven.

Particles, contaminants, and catalyst poisons are among a number of undesired components present in the syngas. Gas cleaning is a required process to remove these undesired contaminants that can cause catalyst deactivation. As shown in Figure 12, gas cleaning typically precedes catalytic synthesis. Large particles (about 5 m) are generally removed with cyclone separators. Smaller particle removal can be accomplished with venturi scrubbers, electrostatic precipitators, and fabric filters. The presence of certain impurities in extremely small concentrations can cause poisoning, reducing the effectiveness of the catalyst. Various processes have been proposed that target specific gas phase contaminants. Sulfur is particularly targeted due to its significant catalyst poisoning activity. A listing of poisons related to methanol synthesis is found in Table 2.

Catalysts are most commonly used in thermochemical processes to convert the synthetic gas into liquid transportation fuels. Commercial considerations require the use of catalysts to alter the reaction kinetics in such a way that the desired yield rate is viable. There is considerable amount of active research focused on increasing catalyst activity and reliability. The choice of catalyst depends on the desired product, operating conditions (temperature and pressure particularly), and economic considerations among others.

Table 2. Methanol Catalyst Poisons and Tolerance Levels

Poison Tolerance Level Reference Sulfur 0.5 ppmv [58]

Halides 0.001 ppmv [59]Iron and Nickel 0.005 ppmv [58]

4.2 Hydrogen Production

The production of hydrogen is generally done through catalytic steam reforming. Biomass conversion to hydrogen has been demonstrated to be economical by Spath et. al.60. Nevertheless, conversion of hydrocarbons from solid feedstock involves higher processing difficulty and

12

capital costs than light hydrocarbons. The representative reaction involves a methane and steam conversion to hydrogen and carbon monoxide.

CH4 + H2O CO + 3H2 (10)

Steam reforming requires heat addition and can be catalyzed over a nickel catalyst. Gasification to produce hydrogen is usually carried under high pressure and temperature conditions. The high pressure is especially necessary when the produced hydrogen serves as feedstock for a subsequent process. Biomass feeding to high pressure systems is an expensive requirement which if resolved could lower plant costs considerably.

Once hydrogen has been generated, it is separated from the synthesis gas. Pressure swing adsorption is a recent technology which involves selective gas to solid adsorption. With this technology, CO2 and H2O are adsorbed by a bed of activated carbon. A second bed with a zeolite molecular sieve further purifies the gas stream. High hydrogen purities of up to 99.99% can be achieved through recycling of the product stream. Ceramic membranes are another advanced technology capable of both gas separation and shift that show great promise.

4.3 Methanol from Syngas

Methanol synthesis development correlates closely with Fischer-Tropsch research. Methanol is currently a commercial product produced mostly from natural gas61. Methanol is formed by the exothermic reaction of one mole of carbon monoxide with two moles of hydrogen:62

CO + 2H2 CH3OH (11)

Low temperatures and high pressures thermodynamically favor the production of methanol. Current commercial operations use a fixed catalytic bed operated at 250 C and 60 - 100 atmospheres with gas recycle to remove the large amount of heat released by this exothermic reaction. More recently, liquid-phase slurry reactors have been introduced to improve contact between syngas and catalyst as well as enhance the removal of heat from the reactor63.

Gas-phase methanol leaving the reactor is mixed with water and higher alcohols. To separate the methanol, a distillation plant is employed where water and volatiles are collected. The remaining syngas is recycled to the reactor increasing the conversion efficiency significantly. Liquid-phase methanol synthesis could reduce investment costs of methanol plants64.

4.4 Mixed Alcohols from Syngas

Efforts in Germany during World War II to develop alternative motor fuels discovered that iron-based catalysts could yield appreciable quantities of water-soluble alcohols from syngas, especially ethanol:65

CO + 3H2 CH3CH2OH (12)

13

These early efforts yielded liquids containing as much as 45-60% alcohols of which 60-70% was ethanol. Working at pressures of around 50 bar and temperatures in the range of 220-370 C,researchers have developed catalysts with selectivity to alcohols of over 95%, but production of pure ethanol has been elusive.

Some researchers have advocated the use of “mixed alcohols” as transportation fuels because the product typically contains a mixture methanol, ethanol, 1-propanol, and 2-propanol. One advantage is the ability to use lower H2:CO ratios than is required for methanol or Fischer-Tropsch synthesis:66

nCO + 2nH2 CnH2n+1OH + (n-1)H2O (13)

with n typically ranging from 1 to 8. The process was commercialized in Germany between 1935 and 1945 but eventually abandoned because of the increased availability of inexpensive petroleum. Working at pressures of around 50 bar and temperatures in the range of 220-370 C,researchers have developed catalysts with selectivity to alcohols of over 95%, but production of pure ethanol has been elusive.

An alternative approach to obtaining neat ethanol from syngas is to first synthesize methanol and subsequently reacting this product with additional syngas:67

CH3OH + 2CO + H2 CH3CH2OH + CO2 (14)

Direct carbonylation of methanol has the advantage of yielding ethanol without coproduct water, which would eliminate energy-intensive distillations. The cost-effectiveness of this approach to ethanol synthesis has not been proven.

4.5 Fischer-Tropsch Synthesis

The biomass to Fischer-Tropsch fuel is a catalyst-based process with precedents in the coal-to-liquids industry. FT catalysts produce hydrocarbons of various lengths of which a significant fraction can be processed into high quality diesel products. Tijmensen et al.68 modeled various configurations for the conversion of biomass to FT liquids. Fischer-Tropsch synthesis produces a large variety of hydrocarbons including light hydrocarbon gases, paraffinic waxes, and alcohols according to the generalized reaction:68

CO + 2H2 —CH2— + H2O (15)

Product distribution is a function of temperature, pressure, feed gas composition (H2/CO), catalyst type and composition. Depending on the types and quantities of F-T products desired, either low (200–240 °C) or high temperature (300–350 °C) synthesis at pressures ranging between 10 to 40 bar are used. For example, high gasoline yield can be achieved using high process temperatures and an iron catalyst. F-T synthesis requires careful control of the H2/COratio to satisfy the stoichiometry of the synthesis reactions as well as avoid deposition of carbon

14

on the catalysts (coking). An optimal H2/CO ratio of 2:1 is maintained through the water-gas shift reaction.

Cobalt and iron are the main active catalyst ingredients used for Fischer-Tropsch synthesis. Cobalt catalysts last longer than iron catalysts and have improved carbon conversion to products due to reduced CO2 generation. Iron, on the other hand, suffers from increased carbon deposition. Since cobalt catalysts can operate at lower pressures, they are commonly used as an alternative to iron catalysts. The costs for cobalt catalysts are about 230 times more expensive than iron, but this cost is offset by lower operating costs.35 Ruthenium is the most active Fischer-Tropsch catalyst but has a prohibitive cost. Much research has been done on improving catalyst activity, lifetime, and selectivity. Adding promoters, improving catalyst preparation, and using zeolites among other methods, can achieve this.

Fischer-Tropsch catalyst can have limited lifetime due to loss of catalyst activity. Poisoning is a controllable problem that can cause significant catalyst deactivation. Table 3 describes the recommended tolerance levels for various poisons. Sulfur is the main Fischer-Tropsch catalyst poison agent, and it is present in both natural gas and coal. Recommendations for sulfur levels are as low as 60 ppb.71 Other poisons include nitrogen compounds and halides. Moisture in the synthetic gas can also have an inhibiting effect due to its water gas shift activity. The efficiency of poison removal will generally be driven by economical considerations since it is nearly impossible to completely remove all synthetic fuel impurities.

Table 3. Fischer-Tropsch Catalyst Poisons and Tolerance Levels

Poison Tolerance Levels ReferenceSulfur 0.2 ppm,

1 ppmv, 60 ppb

[69][70][71]

Halides 10 ppb [70]Nitrogen 10 ppmv NH3,

0.2 ppmv NOx, 10 ppb HCN

[71]

4.6 Fast Pyrolysis

The first fast pyrolysis experiments date from the 1970s and have developed into multiple reactor designs and process configurations. Fast pyrolysis liquids have attracted attention as an intermediate product for the conversion of biomass to liquid fuels and chemicals72.

Fast pyrolysis is the rapid thermal decomposition of organic compounds in the absence of oxygen to produce liquids, gases, and char73. Liquid yields range between 60 to 75%. Typical process conditions consist of moderate temperatures (425-500 C), short vapor residence times (less than 2 seconds), and rapid quenching of pyrolysis vapors and aerosols74.

15

Depolymerization Levoglucosan

Fast

Char + water

Alkali-catalyzeddehydration

Slow

Cellulose Hydroxyacetaldehyde

Figure 2 Reaction pathways in fast pyrolysis

Pyrolysis liquid from flash pyrolysis is a low viscosity, dark-brown fluid with up to 15 to 20% water, which contrasts with the black, tarry liquid resulting from slow pyrolysis or gasification. Typical product yields for two kinds of stored biomass feedstock are given in Table 4. These include acids, aldehydes, sugars, and furans, derived from the carbohydrate fraction, and phenolic compounds, aromatic acids, and aldehydes, derived from the lignin fraction.

Table 4 Material Balance for Fluidized Bed Pyrolysis of Stored Biomass

Material Balance Switchgrass Corn Stover

Gas (%) 11.0 15.4

Char/ash (%) 20.3 15.9

Water + organics (%) 61.5 52.7

Total 92.8 90.9

Gases

CO 3.4 3.7

CO2 7.1 9.9

CH4 0.2 0.4

C2H4 na 1.2

C4H6 0.3 na

The liquid, despite its high water content, shows no appreciable phase separation. However, if an equal volume of water is added to the liquid, the high molecular weight, largely aromatic compounds are precipitated. Since most of the aromatic compounds can be traced to the lignin content of the biomass, this precipitate is widely known as “pyrolytic lignin.” The low pH of pyrolysis liquids, which arises from organic acids derived primarily from the hemicellulosic content of the feedstock, makes the liquid highly corrosive. The liquid also contains fine

16

particulate char. The higher heating values of pyrolysis liquids range between 17 and 20 MJ/kg with liquid densities of about 1280 kg/m3. Assuming conversion of 72% of the biomass feedstock to liquid on a weight basis, yield of pyrolysis oil is about 560 L/ton.

Recovery of high-value chemicals is another possibility, suggesting an integrated approach to production of both chemicals and fuel. For example, levoglucosan is obtained at high yields upon fast pyrolysis of pure cellulose or starch. Even woody or herbaceous biomass can yield significant levoglucosan if metal ions, particularly potassium, are washed from the material before it is pyrolyzed. Levoglucosan is considered a potential building block for synthesis of dextrin-like polymers, pharmaceuticals, pesticides, and surfactants. Microorganisms have been identified that can ferment levoglucosan to citric acid and itaconic acid, which may be an attractive source of these chemicals if the levoglucosan is obtained from inexpensive lignocellulosic materials.

4.7 Bio-oil Gasification

Bio-oil gasification of fast pyrolysis slurries is one possible application for fast pyrolysis liquids with various notable advantages. Gasification of pyrolysis slurries allows for further conversion of pyrolysis products to value-added products such as chemicals and synthetic fuels.

The liquid, char, and tar products from fast pyrolysis can be mixed into a viscous slurry. Pyrolysis slurries can contain up to 85% of the input feedstock’s mass, and an even higher portion of the feed’s energy. Pyrolysis char can be highly reactive and requires careful handling. Mixing the char with pyrolysis oil to form a slurry allows for much easier handling and transportation. This slurry appears almost as a solid. Even at low char concentrations, the mixture loses much of its liquid characteristics because the char absorbs some of the tar and pyrolysis liquids. At 20% straw char and 80% pyrolysis liquids, the mixture can be described as crumbly, thick, and wet bulk material. Various pretreatment measures are necessary prior to feeding the slurry into the gasifier. Atomization of the particles is important for complete carbon conversion with low residence times. Char also has a tendency to agglomerate, and further mechanical means such as using a colloid mixer may be required to break apart the char.

Figure 3 consists of a schematic for a biomass to liquid concept via fast pyrolysis and gasification to Fischer-Tropsch synthesis. As shown, a pyrolyzer unit converts fibrous biomass to gas with some entrained char. Separators, such as cyclones and electro-static precipitators, retrieve the char from the gas. The pyrolysis gas is condensed, optionally at different temperatures, to produce various bio oil fractions in similar manner to petroleum distillation. The char and oil can then be combined as mentioned earlier into a slurry which can be pumped to the gasifier. Once gasified, the synthetic gas generated can be catalytically converted to green diesel fuel.

Bio-oil gasification experiments have been carried out in Germany by Forschungszentrum Karlsruhe75 in collaboration with Future Energy76. Their entrained-flow gasification plant has 350 to 500 kg/h capacity and is configured at 26 bars of pressure. These experiments used straw as feedstock. Slurries with different compositions were used consisting of various pyrolysis

17

liquids and char particle sizes. Char particle sizes ranged from 6 to 50 m, and did not seem to affect the process. Carbon conversion rates of more than 99.5% at 1400 °C were estimated. Viscosities of 2 Pa s and up to 39% solid content were sampled. Problems associated with plugging of nozzles or pipe sedimentation were not reported.

Conversion efficiencies for the pilot plant ranged between 50 and 71%, which are much lower than the theoretical value, but may be due to heat losses associated with the gasifier size. Inert gas flows were also higher (9 to 17% N2) than would be expected for a large scale gasifier. It is estimated that a ton of synthetic fuel can be produced for every 7.5 tons of biomass. Electricity can be generated from the gasification heat to provide energy required for the biomass pretreatment as well as generation and compression of oxygen in the plant.

Figure 3 Bio-oil Gasification

Pyr

olyz

er

Bio-Oil Recovery

Slurry Preparation

Pump Ent

rain

ed F

low

G

asifi

er

Fibrousbiomass

Bio-oil vapor

Slag

Cyclone

Bio-Oil

Char

Fisc

her T

rops

ch

Rea

ctor

GreenDiesel

Pyr

olyz

er

Bio-Oil Recovery

Slurry Preparation

Pump Ent

rain

ed F

low

G

asifi

er

Fibrousbiomass

Bio-oil vapor

Slag

Cyclone

Bio-Oil

Char

Fisc

her T

rops

ch

Rea

ctor

Fisc

her T

rops

ch

Rea

ctor

GreenDiesel

4.8 Bio-Oil Hydrocracking

Bio-oil from fast pyrolysis has superficial resemblance to crude petroleum but its chemical composition is dominated by oxygenated organic compounds rather than hydrocarbons. Thus, efforts to upgrade bio-oil into motor fuel must recognize the considerable mass and energy losses associated with removing oxygen. Early efforts focused on upgrading hydrotreated whole bio-oil at high pressures (2000-2500 psi) and low velocities (0.1 – 0.2 LHSV) to yield an intermediate product that was cracked in an FCC or hydrocracker to produce gasoline.77 Under these circumstances hydrodeoxygenation predominated, requiring large quantities of hydrogen to remove oxygen and generating large amounts of water.

A more attractive approach to upgrading bio-oil is illustrated in Figure 4.78 Whole bio-oil consists of a carbohydrate-derived aqueous phase and a lignin phase. The lignin phase has lower oxygen content than the carbohydrate-derived phase, making it more attractive for hydrotreating. The carbohydrate-derived phase can be steam reformed to provide the hydrogen required for hydrotreating and hydrocracking. Hydrogen consumption could further be reduced by

18

hydrotreating at lower pressures, which favors decarboxylation as the mechanism for oxygen removal.

Hydrotreating tests conducted by UOP LLC and Pacific Northwest National Laboratories78 on pyrolytic lignin are summarized in Table 5. Acid number of the hydrotreated samples dropped to 15 – 30 compared to 30-150 for the raw pyrolytic lignin. Oxygen content dropped from 30% to 6-20%. Further tests were conducted on steam reforming the aqueous phase of bio-oil into oil. Only a small portion of hydrogen was needed to hydrotreat the lignin fraction of the bio-oil; the rest could be used to provide the hydrogen demand for a petroleum hydrocracking at a refinery that was processing both bio-oil and petroleum. The economics of using this biomass-derived hydrogen become attractive when carbon dioxide credits reach $30/ton and bio-oil sells for $10/bbl.

Pyr

olyz

er Carbohydrate derived aqueous phase

Bio-OilRecovery

PhaseSeparation

Steam Reformer

Hyd

rocr

acke

r

Fibrous biomass

Bio-oil vapor

Hydrogen

Dieselfuel

Cyclone

Lignin

Char

Pyr

olyz

er Carbohydrate derived aqueous phase

Bio-OilRecovery

PhaseSeparation

Steam Reformer

Hyd

rocr

acke

r

Fibrous biomass

Bio-oil vapor

Hydrogen

Dieselfuel

Cyclone

Lignin

Char

Figure 4 Bio-oil Hydrocracking

Table 5 Results of hydrocracking pyrolytic lignin78

PNNL UOP WHSV * 0.52 1.0LHSV ** 0.22 0.68Catalyst Pd/C NiMoPressure 1900 – 2000 1500

Liquid yield % 55.6 40.8% Oxygen removal 69 93

% Oxygen in product 19.5 5.9Acid number of product 34 15

% Naphtha in liquid 30 60*Weight hourly space velocity **Liquid hourly space velocity

19

4.9 Hydrothermal Processing

Hydrothermal processing (HTP) describes the thermal treatment of wet biomass to produce carbohydrate, liquid hydrocarbons, or gaseous products depending upon the reaction conditions.79,80 As the desired reaction temperature increases, higher pressures are required to prevent boiling of the water in the wet biomass. Thus, processing conditions range from hot, compressed water at 200 ºC to supercritical water above 374 °C.

At 200 °C and 20 bar pressure, fibrous biomass undergoes a fractionation process to yield cellulose, lignin, and pentose (from hemicellulose). This low temperature HTP has potential application as a pretreatment for enzymatic hydrolysis of cellulose.

At 300 – 350 °C and 120 – 180 bar pressure and reaction times of 5-20 minutes, biomass undergoes more extensive chemical reaction. Reaction products include a hydrocarbon-rich fraction known as biocrude representing as much as 45 wt-% of the product. Products also include 25 wt-% non-condensable gas, which is mostly CO2, and 30% aqueous phase which is about 30 wt-% dissolved organic compounds such as acetic acid and ethanol. The biocrude is attractive for its low oxygen content (10-18 wt-% compared to 45 wt-% for the original biomass) and high lower heating value (30 – 35 MJ/kg). A number of companies have attempted to commercially develop the technology including Biofuel B. V., Changing World Technologies, EnerTech Environmental, Inc., Shell, and TNO.

At 600- 650 °C and 300 bar pressures, supercritical conditions prevail. For reaction times of 0.5-2.0 minutes, gas is the primary reaction product and the process is referred to as supercritical gasification. Typical volume fractions of gas constituents are: H2 – 56 vol-%; CH4 – 47 vol-%; CO2 – 33 vol-%; and CO – 4 vol-%. The combination of relatively low temperature and high pressure makes high methane yields, which is problematic for Fischer-Tropsch synthesis and steam reforming of the gas seems inevitable.

4.10 Catalytic Processing to Methylated Furans

Methylated furans have properties that are potentially attractive as transportation fuels. A schematic of various methylated furans relevant to transportation fuels is shown in Figure 5.They can be produced from both the hexose and pentose content of lignocellulosic biomass and other carbohydrate-rich plant materials. Treatment with acid and/or heat can drive dehydration reactions that ultimately yield furans. Catalysts can improve yields by making furan-producing pathways more selective among the large number of competing reactions that can occur during pyrolysis of biomass.

Sugars consisting of five-member rings like xylose (a pentose) or fructose (a ketohexose) are readily dehydrated to the five-member rings of furan compounds.81 Dehydration of xylose yields furfural (an aldehyde) while fructose yields 5-hydroxymethylfurfural (a dimethylated furan). All described methods of the synthesis of 5-hydroxymethylfurfural require thermal dehydration in acidic medium, which presents some challenges in isolation of this unstable compound. A diagram of proposed reaction pathways to hydroxymethylfurfural is shown in Figure 6.

20

Aldohexoconsiderawhile fru

oses like gluable more di

uctose is obta

ucose can alsifficulty. Xyained from s

so be dehydrylose is obtasucrose or th

rated to 5-hyined from th

he isomerizat

ydroxymethyhe hemi-cellution of gluco

ylfurfural buulose contenose.

ut with nt of biomasss

Figur

Neither fupgradedmethyl fu

Figure(HMF):

Dumesican acid cto a low-conversiohave impThe HMFtwo oxyg

re 5 Furans

furfural nor 5d. Furfural curan, dimeth

6 Proposed: fructopyra

and coworkcatalyst are dboiling-poinon is 77% wproved upon F is next congen atoms lo

relevant to

5-hydroxymcan be methyhyl furan, trim

d reaction paanose (1), fr

kers use a bipdissolved in tnt solvent (m

with half the Hprevious HM

nverted to Dowering the b

the producproces

methylfurfuraylated by reamethyl furan

athway for ructofurano

4) an

phasic reactothe water ph

methyl isobutHMF endingMF synthesiMF using a boiling point

tion of transsing of suga

al is suitable action with mn, and tetram

conversionose (2), two ind HMF (5)4

or to converthase of the retyl ketone).g up in the ois by increascopper-baset to a range t

nsportation fars

as motor fuemethanol ovemethyl furan.

of fructose intermediat49

t fructose to eactor. HMFIn an optimrganic phaseing yields an

ed catalyst. Tthat makes it

fuels by the

el and must ber zeolite cat.82, 83

to hydroxymte stages of d

HMF.49 ThF is extracted

mized system,e. Dumesic and improvinThis converst suitable as

ermochemiccal

be further talyst to yielld

methylfurfudehydration

uraln (3,

he fructose ad as it is form, fructose and coworkeg extractionsion removemotor fuel.

ndmed

ers.s

21

Zhao and coworkers have developed an alternative approach for converting sugars into HMF. Instead of an aqueous solution and acid catalyst, they use ionic liquids as solvent and metal chlorides as catalysts. The absence of an aqueous phase prevents secondary reactions that introduce contaminants to the desired product. The process also works well on both fructose (90% yield) and glucose (70% yield).50

5 Hybrid Processing

Hybrid thermochemical/biological processing thermolytically decomposes fibrous plant materials into a uniform intermediate product that can be biologically converted into a biobased product. There are two distinct approaches to hybrid processing: (1) gasification followed by fermentation of the resulting gaseous mixture of carbon monoxide (CO), hydrogen (H2) and carbon dioxide (CO2), and (2) fast pyrolysis followed by hydrolysis and/or fermentation of the anhydrosugars found in the resulting bio-oil.

5.1 Syngas Fermentation

A number of microorganisms are able to utilize the gaseous constituents of syngas (or producer gas) as substrates for growth and production. These include autotrophs, which use C1compounds as their sole source of carbon and hydrogen as their energy source, and unicarbonotrophs, which use C1 compounds as their sole source of both carbon and energy. Clean syngas from a gasifier can therefore be fed into a stirred fermentor to produce various biobased fuels as shown in Figure 7. Among the fermentation products are carboxylic acids, alcohols, esters, and hydrogen.

In a comprehensive review on the prospects for ethanol from cellulosic biomass, Lynd noted that syngas fermentation represents an “end run” with respect to acid or enzymatic hydrolysis of biomass since it avoids the costly and complicated steps of extracting monosaccharide from lignocellulose.84 It also has the potential for being more energy efficient since it effectively utilizes all the constituents of the feedstock, whether cellulose, hemicellulose, lignin, starch, oil, or protein.

Syngas fermentation also has advantages compared to the use of inorganic catalyst in the production of synthetic fuels.85 Most catalysts used in the petrochemical industry are readily poisoned by sulphur-bearing gases whereas syngas-consuming anaerobes are sulphur tolerant. In conventional catalytic processing, the CO/H2 ratio of the syngas is critical to commercial operations whereas biological catalysts are not sensitive to this ratio; indeed, the water-gas shift reaction is implicit in the metabolism of autotrophic and unicarbonotrophic anaerobes. Gas-phase catalysts typically employ temperatures of several hundreds of degree Centigrade and at least 10 atmospheres of pressure whereas syngas fermentation proceeds at near ambient conditions. Finally, biological catalysts tend to be more product specific than inorganic catalysts.

22

Gasifier

SyngasGas Cleaning

Oxygen

FiberFermenter

Biobasedfuels

CO2

Gasifier

SyngasGas Cleaning

Oxygen

FiberFermenter

Biobasedfuels

CO2

Figure 7 Gasification to biofuels via syngas fermentation

Nevertheless, as described by Grethlein and Jain, syngas fermentation has several barriers to overcome before it can be commercialized.85 Among these are relatively low rates of growth and production by anaerobes, difficulties in maintaining anaerobic fermentations, product inhibition by acids and alcohols, and difficulties in transferring relatively insoluble CO and H2 from the gas phase to the liquid phase, where the anaerobes can utilize the gas. Of these, mass-transfer limitations probably represent the main bottleneck to commercializing this technology. However, studies by Worden and coworkers give encouragement that the use of non-toxic surfactants and novel dispersion devices can enhance mass transfer through the generation of microbubbles to carry syngas into bioreactors.86

5.2 Bio-Oil Fermentation

Bio-oil fermentation is based on the observation that under rapid pyrolytic conditions, pure cellulose yields levoglucosan, an anhydrosugar with the same empirical formula as the monomeric building block of cellulose: C6H10O5.87 Anhydrosugar is a sugar from which one or more molecules of water have been removed, resulting in the formation of an internal acetal structure as shown in Figure 8.

On the other hand, addition of a small amount of alkali inhibits the formation of levoglucosan and promotes the formation of hydroxyacetaldehyde; for slower heating rates and lower temperatures char rather than liquids is preferentially formed. These multiple reaction pathways for pyrolysis of cellulose are illustrated in Figure 2.88

Scott and coworkers at the University of Waterloo in Ontario, Canada recognized that alkali in biomass catalyzed the char-forming pathway.89 If these cations are removed by soaking the feedstock in dilute acid before pyrolysis, the lignocellulose is depolymerized to anhydrosugars, primarily levoglucosan, at very high yields. Levoglucosan is readily hydrolyzed to glucose.

23

Brown and his collaborators evaluated the effect of alkali removal on the pyrolytic products of cornstover.90 Three pretreatments were evaluated: acid hydrolysis, washing in dilute nitric acid, and washing in dilute nitric acid with the addition of (NH4)2SO4 as a pyrolytic catalyst. All three acid treatments were able to substantially increase the yield of anhydrosugars from less than 3 wt-% to as high as 28 wt-%. Acid hydrolysis of this anhydrosugar yielded 5% solutions of glucose and other simple sugars.

The resulting glucose solutions can be fermented, as demonstrated by Prosen et al.91 However, the resulting substrate derived from the bio-oil contains fermentation inhibitors that must be removed or neutralized by chemical or biological methods. Chemical methods that have been evaluated on bio-oil derived substrate include solvent extraction, hydrophilic extraction, and adsorption extraction92. Khiyami explored the use of ligninolytic enzymes to remove toxins from bio-oil93.

One manifestation of a biorefinery based on fermentation of bio-oil is illustrated in Figure 9. Fibrous biomass is pretreated with dilute acid to simultaneously remove alkali and hydrolyzes the hemicellulose fraction to pentose. The remaining fraction, containing cellulose and lignin, is pyrolyzed at 500 C to yield char, gas, and bio-oil. The bio-oil is separated into pyrolytic lignin and levoglucosan-rich aqueous phase. The char, gas, and lignin are burned to generate steam for distillation and other process heat requirements of the plant while the levoglucosan is hydrolyzed to hexose. The pentose and hexose are fermented to ethanol.

Figure 8 Chemical structure for levoglucosan (1,6-anhydro-beta-D-glucose)

24

Figure 9 Bio-oil Fermentation

6 Economics of Biorefineries

6.1 Biochemical Refineries

Representative of biochemical refineries is the corn grain to ethanol process. Fuel ethanol is commercially available as a transportation fuel or additive. Most fuel ethanol production is derived of fermentation of corn glucose or sucrose in the US94 and sugarcane in Brazil95

respectively. Dry milling and wet milling are general classifications for the processing of corn-based ethanol. Due to lower requirements for capital and labor, dry milling has dominated the recent grain ethanol industry. Advanced biochemical refineries convert cellulose from plant fibers to ethanol. Ethanol from lignocellulosic biomass employs hydrolysis, or other pre-treatments, to break down the cellulosic material prior to fermentation. Enzymatic processes are expected to have much lower costs in the future than current starch based processes96.

Fermenter

Fiber

Pyr

olyz

er

Anhydrosugar & other carbohydrate

Bio-OilRecovery

PhaseSeparation

Detoxification

Lignin

Hot water extraction

Pentose

Fiberbyproduct

Bio-oil vapor

Fermenter

Distillation

Water

Ethanol

Cyclone

CharDrie

r

Fermenter

Fiber

Pyr

olyz

er

Anhydrosugar & other carbohydrate

Bio-OilRecovery

PhaseSeparation

Detoxification

Lignin

Hot water extraction

Pentose

Fiberbyproduct

Bio-oil vapor

Fermenter

Distillation

Water

Ethanol

Cyclone

Drie

r

Char

25

6.1.1 Grain Ethanol

Starch-based ethanol fermentation plants are relatively inexpensive to build and operate. The general operation of the plant follows the schematic outlined in Figure 10. Capital costs for 25 million gallons of ethanol capacity can be as low as $27.9 million97. Production cost estimates in the USA and Europe range between $0.9698 and $1.3799 per gallon depending on assumptions, especially feedstock costs. McAloon et. al. estimated the operating costs of a 25 million gallon per year capacity plant to total $1.00 (2005 dollars) per gallon of ethanol. Table 6 consists of the operating cost components of this plant.

consists of the operating cost components of this plant.

The purchase of corn grain is the greatest expense for fermentation plants. Corn grain prices have historically been about $2 per bushel, but current trends find the price upwards of $3 per bushel. Distillation is an energy intensive process that adds considerable cost. The fermentation process yields a mixture of mostly water and 4-5% by weight ethanol. Distillation followed by filtration through vapor-phase molecular sieves is required to achieve 99.5% ethanol purity. This process also produces various by-products such as distiller’s dried grain (DDGS). Revenue from DDGS can offset operation and management costs.

The purchase of corn grain is the greatest expense for fermentation plants. Corn grain prices have historically been about $2 per bushel, but current trends find the price upwards of $3 per bushel. Distillation is an energy intensive process that adds considerable cost. The fermentation process yields a mixture of mostly water and 4-5% by weight ethanol. Distillation followed by filtration through vapor-phase molecular sieves is required to achieve 99.5% ethanol purity. This process also produces various by-products such as distiller’s dried grain (DDGS). Revenue from DDGS can offset operation and management costs.

Table 6 Corn Grain Production Costs (McAloon et. al. 1999) Table 6 Corn Grain Production Costs (McAloon et. al. 1999)

AnnualAnnual Per GallonPer GallonShelled Corn $17,000,000 $0.68Other Raw Materials $1,600,000 $0.06Denaturant $600,000 $0.03Utilities $4,000,000 $0.16Labor, Supplies, and Overhead Expenses $3,100,000 $0.13Depreciation of Capital $2,800,000 $0.11DDG Credit -$7,100,000 -$0.29Total Production Cost $22,000,000 $0.88

CO2StarchEnzymes

Fermenter

Grain Cooking

StillageDrying

Distillation

Grinding

Ethanol

DistillersDried

Grains

Figure 10 Grain-to-ethanol processing

26

6.1.2 CCellulosic Ethanol

Advancethe corn estimatescapital cowith adva

The biocethanol penzymesto productechnologannual caupon the be emploProjectedbe $3.33 $0.89 per

Processinrequires mcarbohydresidencerecycledby-produ

F

d biochemicstalk that is s at $50 per Mosts remain hanced techno

hemical platplatform in th; enzymatic ce ethanol; agy, capital capacity. Opeassumption

oyed. A reprd capital cost- $4.44 per

r gallon of e

ng of cellulomore intensi

drates requiree time. The mand used for

ucts, such as

Figure 11 C

cal refineriestypically notMg. While fhigher due toology as hig

tform based he unit operahydrolysis o

and distillatioosts are estim

erating costs s made abouresentative st for future pgallon ethanthanol.103

osic biomass ive pretreatmes additionamain by-prodr process heorganics.

Cellulosic Et

s use cellulost harvested, feedstock coo added proc

gher process

on enzymatations involvof cellulose ton to producmated to be are estimate

ut feedstock schematic ofplants emplonol annual ca

differs fromment dependl energy, andduct from ceat. Addition

thanol Oper

sic biomasshas a much sts are lowercess complexefficiencies

ic hydrolysived: pretreatto sugars; fece neat ethanbetween $4.

ed to be betwcosts, enzym

f operating cooying more aapacity with

m corn grain ding on how d the fermenellulosic biomal revenue m

rating Costs

as feedstocklower cost thr for the advxity. Operatof up to 48%

s of cellulostment to makrmentation o

nol. Based o.03 and $5.6ween $1.34 1

me costs, andosts componadvanced tecoperating co

in some keyit is delivere

ntation procemass conver

may be deriv

s (50 Million

k. Corn stovehan corn gra

vanced bioching costs can% are achiev

er, a portion ain with currhemical procn be reducedved100.

ofrent cess, d

e is very simke the celluloof sugars by on the curren60 per gallon101 and $1.69d the kind ofnents is showchnologies arosts droppin

milar to the gose accessibyeast or bac

nt state of n of ethanol 9 102 dependf pretreatmenwn in Figurere estimated

ng to $0.40 an

grainble to cteria

ingnt to 11.

d to nd

y areas. Cellued. Reducingess necessitarsion is ligni

ved from othe

ulosic biomag the compleates a longer in, which caner small volu

assex

n be ume

n Gallons peer Year)

27

6.2 Thermochemical Refineries

Unlike the biochemical platform, for which the fuel product is defined (ethanol) and there is reasonable consensus as to the unit operations to be employed, there is a greater diversity of opinion on how the thermochemical platform should be configured. Thermolytic processing of biomass can generally be categorized as either gasification (to produce a gaseous product) or fast pyrolysis (to produce primarily liquid product). A variety of catalytic 104 or even biocatalytic schemes4 have been proposed to upgrade the thermolytic products into alcohols, ethers, esters, or hydrocarbons. A typical catalytic synthesis plant based on biomass gasification can be described by the diagram shown in Figure 12. Thermochemical plants are better suited for large-scale, 100 million to 500 million or more gallons per year, deployment due to high capital costs. Economies of scale allow for lower costs per unit volume, but transportation costs need to be taken into consideration as well as feedstock availability. Based on capital, operating, and transportation costs, an optimum plant size can be determined at which fuel production costs are minimized.

6.2.1 Hydrogen

Thermochemical treatment of biomass to produce hydrogen, or methanol, has a high efficiency of 50%, and 45% respectively. The implication of this is that for the same energy input hydrogen plants can yield a greater amount of fuel energy than corn grain ethanol for example. This is reflected in a lower feedstock cost due to a lower feed input requirement. Operation and management costs are small, and in some cases, completely accounted for through by-product revenues. Hydrogen plants can be configured to produce heat or electricity from excess methanol production.

The literature includes several studies on the cost of hydrogen from biomass. Capital costs range from $0.65 to $1.33 per gallon of hydrogen capacity105, 106, depending upon the type and size of gasifier plant. Operating costs range from $0.31 to $0.44 per gallon of hydrogen produced 107

depending upon the cost of biomass and the kinds of processes employed108. Hamelinck and Faaij 109 represent one of the most recent studies (2002) on biomass to hydrogen. For a gasification plant producing 182 million gallons hydrogen fuel per year, the estimated capital

Catalytic Synthesis

Gasifier

SyngasGas Cleaning

Biobased fuels

CO2

Oxygen

ber Fi

Figure 12 Gasification to fuels via catalytic synthesis

28

cost is $2includes

206 MUS withe major op

ith operatingperating cost

g costs of $0t component

.24 per gallots for a hydro

on hydrogenogen plant.

n produced. FFigure 13

Figurre 13 Annuaal operatingg costs for 4400 MW hyddrogen plannt

6.2.2 MMethanol SSynthesiss

ProductioPreviousfrom $0.6and FaaijgasificatiinvestmeOperatin

on costs for techno-econ

62 to $1.11 pj analyzes boion plant proent of $276 Mg cost comp

methanol arnomic studieper gallon ofoth hydrogenoducing 87 mMUS. The pronents are sh

re identical wes of methanf methanol11

n and methamillion galloroduction cohown in Fig

within a 30%nol plants ind10, 111. The tnol plants. Bns of methan

ost of methangure 14.

% uncertaintydicate that mtechno-econoBased on avenol per year nol was $0.6

y to those ofmethanol costomic analysierages from would requi

62 per gallon

f a hydrogen ts would ranis of Hamelitheir analysiire a capital

n of methano

plant. ngeinckis, a

ol.

Hamelincproductioprocessinspecific ipressurerefining aimportanthan hydrhydrogen

ck and Faaijon of methanng sections ainvestments swing adsorand make up

nt consideratirogen at 16 Mn is required

modeled thnol and hydrare interchancosts related

rption units ap compressoion is the enMJ/L (compto provide t

e costs of varogen. Whilengeable for ed to the geneat a cost of ar add up to $

nergy contentpared to 8 Mthe same am

arious therme the gasificaeither hydrogeration of eacabout $30 mi$30 million ft of the fuel.J/L hydroge

mount of ener

mochemical cation, gas clegen or methach fuel. Hydillion for a 4for an equiv. Methanol hen), thereforergy.

configurationeaning, and anol productdrogen produ400 MW planalent plant. A

has twice thee a larger vo

ns for the syngastion, there aruction requirnt. MethanoAnothere energy denlume of

reresl

nsity

29

6.2.3 M

In April owith hydnatural gmixed alc

Some eco13% for Mtechnologcost of met. al. sugethanol b

Feed

P

Retu

Figu

Mixed Alco

of 2005, therdrocarbon ch

as due to ecocohols.

onomic studMTBE fromgy develope

mixed alcohoggest that ethby 2012116.

T

Product(s)d Rate (Pet C

Feed Cost Plant Cost

Product Valu

urn on Invest

re 14 Annu

ohols

re were no cains of six conomical rea

dy results arem C1 to C4 alc

rs include Dols from othehanol and m

Table 7 Mix

Coke)

ue

tment

al operating

commercial parbons or leasons. Synga

e included incohols, and

Dow113, Snamer references

mixed alcohol

xed Alcohol

BMTBE fro

1

$35

$0.85

g costs for 4

plants dedicass112. Mixedas from biom

n Table 7. Th15% for etha

mprogetti (SEs is shown inl synthesis c

Economics

Bechtel117

om C1/C4 al800 TPD $0/ton

50-400MM

5 cents/gal M

~13%

400 MW me

ated to the pd alcohols armass can be

he expected ranol, propanEHT)114, an

n Table 8. Ccould be cost

s Study Com

cohols E

MTBE

ethanol plan

production ofre generally pused to cata

return on invnol, and powd IFP (Substonclusions mt competitive

mparison

Fluor/EEthanol, Prop

475$0

estimated at$1.15

alcohols

~

nt

f mixed alcoproduced frolytically pro

oholsomoduce

vestment is awer. Other

tifuel)115. Thmade by Phie with corn

about

hellips

Ecalene118

panol, and Po58 TPD 0/ton t >$1 billioncents/gal fo, 4.4 cents/kpower

~15%

ower

n63orkWh

30

Table 8 Mixed Alcoohol Producction Costs ((adapted from Spath 22003)

Prod(ton

55386073

635,000~1,7

6.2.4 F

Biomass Tropschnatural gCapital cslurry ph

Previouswould racompleteal. 68 Theabout 35 about $2

F

ductionnne/yr)0,0006,0000,0000,0000-730,000

700,000

Fischer-Tr

to Fischer-Tcommercial as or coal ad

cost estimatehase reactors

studies of bange from $1e techno-ecoey estimate tmillion gall

.37 per gallo

Figure 15 An

ropsch Sy

Tropsch fuelplants empl

dds additionaes by Sasol fo

equal about

biomass gasif.1 to $4.1 penomic analy

the capital colons of F-T don of F-T die

nnual opera

Metha

5063

(60-70)70

70/3070

ynthesis

studies are loying other al cost and s

for a two trait $25,000 pe

fication planer gallon of Fyses of a Fiscost for an oxdiesel per yeesel. A summ

ating cost fo

nol/C5+

0/50/35

)/(30-40)0/30, 50/50

0/30

generally bafossil fuels.

some complen Gas-To-Li

er BPD capac

nts producingF-T diesel35,

cher-Tropschxygen-blownear to be aboumary of the o

or 367 MW

Cost ($/

$167$255 -

$31$395 -$282 -$318 -

/tonne)

7.7 $198 3.9 $110 $205 $279

ased on the aThe use of b

exity due to piquids plant city125.

g Fischer-Tr, 105. Amongh diesel plan

n, pressurizedut $341 MUoperating co

biomass Fis

already estabbiomass as apretreatmentwith 30,000

ropsch diese the most rec

nt is that by Td gasifier pla

US with operst is shown i

scher-Trops

Cost Year

19861989198319821986

na

r Refer

[11[12[12[12[12[12

ence

9]0]1]2]3]4]

blished Fischa substitute tt considerati

0 barrels per

her-toions.day

l suggest coscent and Tijmensen eant producinating costs oin Figure 15

sts

etngof5.

sch plant

31

Fischer-Tropsch synthesis yields a wide variety of hydrocarbon chains. Small hydrocarbons (C1– C4) can be recycled and used as process heat. Long hydrocarbons can be broken into smaller chains if necessary. The Anderson-Schulz-Flory (ASF) distribution describes the relation between hydrocarbon yield and chain growth126. Commercializing the by-products is important for Fischer-Tropsch biomass diesel to be become economically viable.

6.2.5 Bio-oil Fermentation

So and Brown127 have performed one of the only techno-economic analyses of bio-oil fermentation. As summarized in Table 9, total capital investment for an ethanol plant based on fermentation of bio-oil producing 95 million L of ethanol was estimated to be $69 million, while the annual operating cost was about $39.2 million, resulting in an ethanol production cost of $ 0.42/L.

Table 9 Production cost of ethanol from cellulosic biomass

Annual ethanol output 95 million L Annual biomass input 240 x 106 kg

Total capital $69 million

Raw materials ($46/ton) $11.1 million Labor, utilities, maintenance $6.18 million

Indirect costs $8.07 million Annual capital charges $13.8 million Annual operating costs $39.2 million

Production cost of ethanol $0.42/L 11997 US $

6.2.6 Syngas Fermentation

Although syngas fermentation to ethanol is undergoing commercial development, there are no publications on ethanol production from syngas fermentation to date. We have performed a preliminary economic assessment on syngas fermentation to hydrogen and polyhydroxyalkonate (PHA), a biodegradable polymer of commercial interest.128 A diagram of the process is included in Figure 16.

This analysis assumes 20 tpd of PHA and 55 tpd of hydrogen. The capital costs, detailed in Table 10, are estimated to be $103 million. The operating costs for this plant, shown in Table11, yields a production cost of $2.80/kg of PHA assuming a credit for co-product hydrogen equal to $1.90/kg, which is based on a U.S. Department of Energy long-term target price for this fuel. This PHA production cost compares favorably with the production cost of PHA from glucose, which may be as much as $5-7/kg.

32

Tar Cracker Methane Reformer

Syngas Bioreactor Separations

Gasifier

Figure 16 Conceptual schematic of biorefinery to produce hydrogen and PHA coproducts from fibrous biomass

Table 10 Estimated capital costs for biorefinery to produce hydrogen and PHA coproducts from fibrous biomass

Gasifier $18.6 million Estimated from: Larson and Svenningsson (1990)

Fermenter $59.1 million Estimated as 25% of total cost of an ethanol plant

Separation equipment $25.3 million Estimated as 30% of total costs of a fermentation plant

Grassroots Capital $103 million

33

34

Table 11 Estimated operating costs for biorefinery to produce hydrogen and PHA coproducts from fibrous biomass

Annual H2 output 18.1 x 106 kg Based on 20% CO to cell mass; 35% cell mass to PHA

Annual PHA output 6.5 x 106 kg

Annual input 210 x 106 kg 90% capacity factor

Total Capital $119 million Raw materials $12.6 million Purchased at $0.06/kg Credit for H2 ($34.3 million) Assumed to sell for $1.90/kg

Labor, utilities, maintenance $15.6 million Indirect costs $10.4 million

Annual capital charges $13.9 million 10% interest, 20 yrs Annual operating costs $18.2 million PHA Production costs $2.80 / kg

7 Conclusions

Current generation biofuels will continue to be heavily employed due to the infrastructure and policies already in place. Advanced fuels from biorenewable sources, particularly cellulosic biomass, will be necessary to displace larger amounts of fossil fuel resources. Various technologies stand to profit from the renewed interest in alternative sources of energy.

As shown in this paper, each of the potential biofuels considered has its advantages and disadvantages in terms of market applications, environmental benefits, and process economics. This results in a situation where no single biofuel may ultimately dominate in similar fashion to gasoline. Niche markets and regional considerations will heavily influence the best choice of fuel for specific applications. Finally, increased research efforts could yield raise new possibilities that change the outlook of any, and all, bio-based fuels.

Although biochemical processing is the dominant form of biomass based fuel in the current generation, thermochemical and hybrid processing stand to increase their contributions. A major motivation for the interest in these advanced pathways is the increased in biochemical feed costs due to the high demand and competition with food demand. Cellulosic based technologies enjoy a larger pool of feed sources, including waste materials, and are more flexible in terms of the biomass inputs. Hybrid processing allows for the combination of the thermochemical flexibility with the biochemical established technology.

It is almost certain that the predominant biofuel choice will be based on a combination of market forces, environmental concerns, and public policy drivers. At this point, there may not be enough information to determine a ‘winner’ among the various choices. Nevertheless, increased efforts in this research area should help narrow down the available choices to biorenewable resources that can serve as a foundation for our energy future.

8 References

1 Bodman, Samuel. U.S. Chamger of Commerce: Biofuels and the Future of U.S. Energy. United States Department of Energy Office of Public Affairs. October 25, 2006.

2 Sun, Ye, Cheng, Jiayang (2002) Hydrolysis of lignocellulosic materials for ethanol production: A review”. Bioresource Technology 83, 1-11.

3 Reed, T. ed. 1981. Biomass Gasification: Principles and Technology, Noyes Data Corp., Park Ridge, NJ.

4 R. C. Brown, Biomass Refineries based on Hybrid Thermochemical/Biological Processing– An Overview, in Biorefineries, Eds. Biobased Industrial Processes and Products, Kamm, B., Gruber, P. R., Kamm, M., Wiley-VCH Verlag, Weinheim, Germany, (2005).

5 Pengmei L., Zhenhong Y., Chuangzhi W., Longlong M., Yong C., Noritatsu T. “Bio-syngas production from biomass catalytic gasification” Energy Conversion and Management 48 (2007) 1132-1139

6 Bridgwater, A. V. and G. V. C. Peacocke (2000). Fast pyrolysis processes for biomass, Renewable and Sustainable Energy Reviews 4: 1-73.

7 Marinangel, R., Marker, R., Petri, J., Kalnes, T., McCall, M., Mackowia, D, Jerosky, B., Ragan,B., Nemeth, L., Krawczyk, M., Czernik, S., Elliott, D., Shonnard, D. (2006) Opportunities for Biorenewables in Oil Refineries, Department of Energy Final Technical Report, Contract No. DE-FG36-05GO15085.

8 Elliott, D. C., Sealock, L. J., Jr., Baker, E. G., Butner, R. S. (1993) Chemical Processing in High-Pressure Aqueous Environments. 1. Historical Perspective and Continuing Developments. Ind. Eng. Chem. Res. 32, 1542.

9 Bobleter, O.(1994) Hydrothermal degradation of polymers derived from plants. Prog. Polym. Sci. 19, 797–841.

10 Allen, S.G., Kam, L.C., Zemann, A.J. and Antal, Jr., M.J. (1996) Fractionation of sugar cane with hot, compressed, liquid water. Ind. Eng. Chem. Res. 35, 2709–2715.

11 Goudriaan, F., Peferoen, D. (1990) Liquid fuels from biomass via a hydrothermal process, Chem. Eng. Sci. 45, 2729–2734.

12 Antal, M.J., Allen, S., Schulman, D., Xu, X., and Divilio,R. (2000) Biomass gasification in supercritical water, Ind. Chem. Eng. Res. 39, 4040–4053.

13 Brown, R. C., Chapter 24. Biomass Energy Conversion, Section 24.3 Bio-fuels, CRC Handbook of Energy Conservation and Renewable Energy, Kreith, F. and Goswami, Y., Eds., CRC Press, in press.

14 Krawczyk, T. (1999) Specialty oils, INFORM - International News on Fats, Oils and Related Materials 10(6): 552.

15 Basiron Y. “Palm Oil Production through Sustainable Plantations” European Journal of Lipid Science and Technology (2007) 109, 4, 289-295

16 Sheehan J., Dunahay T. Benemann J., Roessler P. “A Look Back at the U.S. Department of Energy’s Aquatic Species Program: Biodiesel from Algae.” NREL/TP-580-24190 July 1998.

17 Cusano, C. M.; Stafford, R. J.; Lucas, J. M. Changes in elastomer swell with diesel fuel composition. SAE Technical Paper Series 1994, 942017.

18 Bacha, J.; Blondis, L.; et al. Chevron Diesel Fuels Technical Review. 1998, (FTR-2).

35

36

19 Heinze, P. Engine performance and emissions with future type diesel fuels. Presented at the Institution of Mechanical Engineer International Conference on Petroleum Based fuels and Automotive Applications, 1986; Paper No. C306/86.

20 Barbour, R. H.; Rickeard, D. J.; Elliot, N. G. Understanding Diesel Lubricity. SAE Technical Paper Series. 2000, 2000-01-01918.

21 Lamprecht D. “Performance Synergies between Low-Temperatures and High-Temperature Fischer-Tropsch Diesel Blends” Energy & Fuels 2007, 21, 2846-2852

22 Hamelinck C. N. Faaij A. P. C., Uil H. den, Boerrigter H. “Production of FT transportation fuels from biomass technical options, process analysis and optimization and developmental potential. Energy 29 (2004) 1743-1771

23 Bailey, B. K. (1996) Performance of Ethanol as a Transportation Fuel, in Handbook on Bioethanol: Production and Utilization, Wyman, C. E., Taylor & Francis, Washington, DC.

24 Lynd, L. R., Cushman, J.H., Nichols, R. J., and Wyman, C. E. (1991) Fuel Ethanol from Cellulosic Biomass, Science 251, 1318-1323.

25 Klass, D. L. (1998) Biomass for Renewable Energy, Fuels, and Chemicals, Academic Press, San Diego, pp. 401-402.

26 Davenport, B. “Methanol.” Chemical Economics Handbook Marketing Research Report, SRI International. Report Number 674.5000, 2002.

27 Gray Jr., C. L. and Alson, J. A. (1989) The case for methanol, Scientific American 261, 108-114.

28 “Beyond the Internal Combustion Engine” American Methanol Institute. Accessed: June 2007. http://www.methanol.org/fuelcell/special/ami.pdf

29 Durre, P.(2007) Biobutanol: An attractive biofuel, Biotechnology Journal 2, 1-10. 30 Schuegerl, K. “Integrated processing of biotechnology products”. Biotechnology Advances, v

18, n 7, Nov 2000. P 581-599. 31 David Ramey, and Shang-Tian Yang. “Production of Butyric Acid and Butanol from Biomass”

U. S. Department of Energy. Report 2004. 32 Ezeji T. C., Qureshi N., Blaschek H. P. “Bioproduction of Butanol from Biomass: from Genes

to Bioreactors” Current Opinion in Biotechnology18, 3, June 2007 220-227 33 Klass, D. L. (1998) Biomass for Renewable Energy, Fuels, and Chemicals, Academic Press,

San Diego, p. 427-429. 34 Forzatti, P., Tronconi, E., and Pasquon, I. (1991) Higher Alcohol Synthesis, Catalysis

Reviews-Science and Engineering, 33(1-2), 109-168. 35 Spath, P. L. and Dayton, D. C. (2003) Preliminary Screening - Technical and Economic

Assessment of Synthesis Gas to Fuels and Chemicals with Emphasis on the Potential for Biomass-Derived Syngas, National Renewable Energy Laboratory Report NREL/TP-510-34929, December.

36 Fischer-Tropsch Archive (2005) www.fischer-tropsch.org. 37 Schulz H. “Short History and Present Trends of Fischer-Tropsch Synthesis” Applied

Catalysis A: General 186 (1999) 3 - 12 38 Senden, M. M. G., Sie, S. T., Post, M. F. M. and Ansorge, J. (1992). "Engineering

aspects of the conversion of natural gas into middle distillates." NATO ASI Series, Series E: Applied Sciences 225(Chemical Reactor Technology for Environmentally Safe Reactors and Products): 227-47.

37

39 Romm, J. J. (2004) The Hype About Hydrogen: Fact and Fiction in the Race to Save the Climate, Island Press, Washington, D. C.

40 Vieira de Carvalho, A. J. (1982) Natural gas and other alternative fuels for transportation purposes, Energy 10(2): 187-215.

41 Anon (2005) Biogas replaces natural gas for vehicles, BioCycle 46(10): 55 42 Klass, D. L. (1998) “Chapter 12. Microbial Conversion: Gasification,” Biomass for

Renewable Energy, Fuels, and Chemicals, Academic Press, San Diego. 43 Reed, T., editor (1981) Biomass Gasification: Principles and Technology, Noyes Data Corp.,

Park Ridge, N.J. 44 Sorenson, S. C. (2001) Dimethyl Ether in Diesel Engines: Progress and Perspectives, Journal

of Engineering for Gas Turbines and Power 123(3): 652-658. 45 Hansen, J. B., and Joensen, F. (1991). "High conversion of synthesis gas into oxygenates."

Studies in Surface Science and Catalysis, 61(Nat. Gas Convers.), 457-67. 46 Peng, X. D., Toseland, B. A., and Tijm, P. J. A. (1999). "Kinetic understanding of the

chemical synergy under LPDME conditions-once-through applications." Chemical Engineering Science, 54(13-14), 2787-2792.

47 Moreau, et al.Recent Catalytic Advances in the Chemistry of Substituted Furans from Carbohydrates and in the Ensuing Polymers, Top. Catal. Volume 27, 2004. Pages: 11-30.

48 Hanniff et al., Conversion of Biomass Carbohydrates into Hydrocarbon Products, Proceedings of IGT/CBETS Conference on Energy from Biomass and Wastes X, Washington, D.C., Apr. 7-10, 1986.

49 Roman-Leshkov, Barrett, Liu & Dumesic; Production of dimethylfuran for liquid fuels from biomass-derived carbohydrates, Nature, 447, 982-985 (21 June 2007).

50 Zhao, et al, Science, 2007, 316, 1597. 51 Rowell, R. M., Young, R. A., and Rowell, J. R., Paper and Composites from Agro-Based

Resources, CRC Press, 1997. 52 Schell, D., McMillan, J. D., Philippidis, G. P, Hinman, N. D., and Riley, C. (1992) Ethanol

From Lignocellulosic Biomass, in Advances in Solar Energy: Volume 7, K. Boer, Ed., American Solar Energy Society, pp. 373-448.

53 Brown, R. C. (2003) Biorenewable Resources: Engineering New Products from Agriculture, Blackwell Publishing, Ames, IA.

54 Lynd, L. R. (1996) Overview and evaluation of fuel ethanol from cellulosic biomass: Technology, economics, the environment, and policy, Annu. Rev. Energy Environ. 21, 403 - 465.

55 Wayman, M. and Parekh, S. R. (1990) Chapter 16. Ethanol from Wood and Other Cellulosics, in Biotechnology of Biomass Conversion: Fuels and Chemicals from Renewable Resources, Open University Press, Philadelphia.

56 Ronald F. Probstein, and R. Edwin Hicks. “Synthetic Fuels”, Mineola, New York: Dover Publications, 2006.

57 Chem. Eng. News (1982) Technology: Catalyst Converts Methanol to Dry Ethanol 60 (38), 41, September 20.

58 Kung, H. H. (1992). "Deactivation of Methanol Synthesis Catalysts - a Review." Catalysis Today, 11(4), 443-453.

59 Twigg, M. V., and Spencer, M. S. (2001). "Deactivation of supported copper metal catalysts for hydrogenation reactions." Applied Catalysis, A: General, 212(1-2), 161-174.

38

60 Spath P. L., Mann M. K., and Amos W. A. “Update of Hydrogen from Biomass – Determination of the Delivered Cost of Hydrogen.” National Renewable Energy Laboratory Report NREL/MP-510-33112, December 2003.

61 Davenport, B. (2002). "Methanol." Chemical Economics Handbook Marketing Research Report, SRI International, Menlo Park, CA. Report number 674.5000.

62 Beenackers, A. A.C. M. and Van Swaaij, W. P. M. (1984) Methanol from wood: Process principles and technologies for producing methanol from biomass, Int. J. Solar Energy 2, 349-367.

63 A. Cybulski, Liquid phase methanol synthesis: catalysts, mechanism, kinetics, chemical equilibria, vapor–liquid equilibria, and modeling—a review, Catal. Rev. Sci. Eng. 36 (4) (1994) 557–615.

64 USDOE, Commercial-scale Demonstration of the Liquid Phase Methanol (LPMEOHTM) Process—Clean Coal Technology Topical Report #11, US Department of Energy, 1999.

65 Klass, D. L. (1998) Biomass for Renewable Energy, Fuels, and Chemicals, Academic Press, San Diego, p. 427-429.

66 Forzatti, P., Tronconi, E., and Pasquon, I. (1991) Higher Alcohol Synthesis, Catalysis Reviews-Science and Engineering, 33(1-2), 109-168.

67 Chem. Eng. News (1982) Technology: Catalyst Converts Methanol to Dry Ethanol 60 (38), 41, September 20.

68 M. J.A. Tijmensen, A. P.C. Faaij, C. N. Hamelinck, and M. R.M. van Hardeveld, “Exploration of the possibilities for production of Fischer-Tropsch liquids and power via biomass gasification”. Biomass and Bioenergy 23. 129-152, 2002.

69 Dry, M. E. (1999) “Fischer-Tropsch reactions and the environment”, Appl. Catal. A, 189(2): 185-190.

70 Boerrigter, H., den Uil, H., and Calis, H.-P. (2002). "Green Diesel from Biomass via Fischer-Tropsch Synthesis: New Insights in Gas Cleaning and Process Design." Paper presented a Pyrolysis and Gasification of Biomass and waste, Expert Meeting, 30 September, 2002, Strasbourg, FR.

71 Turk, B.S., Merkel, T., Lopez-Ortiz, A., Gupta, R.P. Portzer, J.W., Kishnam, G., Freeman, B.D., and Fleming, G.K. (2001). "Novel TEchnologies for Gaseous Contaminants Control." Final Reprot for DOE Contract No. DE-AC26-99FT40675, September 2001.

72 Bridgwater A. V., Peacocke G.V.C. “Fast Pyrolysis Processes for Biomass” Renewable and Sustainable Energy Reviews 4 (2000) 1-73

73 Bridgwater, A. V. “Production of high grade fuels and chemicals from catalytic pyrolysis of biomass” Catalysis Today 1996, 29 (1-4), 285-295.

74 Bridgwater, A. V. “Renewable fuels and chemicals by thermal processing of biomass” Chemical Engineering Journal 2003, 91, 87-102.

75 Klaus Raffelt, Edmund Henrich, Andrea Koegel, Ralph Stahl, Joachim Steinhardt, and Friedhelm Weirich. “The BTL2 Process of Biomass Utilization Entrained-Flow Gasification of Pyrolyzed Biomass Slurries.” Applied Biochemistry and Biotechnology 129. 153 – 164.

76 Future Energy GmbH, Halsbrücker Stra e 34, 09599 Freiberg, Germany, www.future-energy.de. Date accessed: May 2005.

39

77 Grange, P, Laureat, E.,Maggi, R, Centeno,A., Delmon, B (1996) Hydrotreatment of pyrolysis oil from biomass: Reactivity of the various categories of oxgenated compounds and preliminary technico-economic study, Catalysis Today 29 297-301.

78 Marker, T. L. (2005) Opportunities for biorenewables in oil refineries, Department Of Energy Technical Report, Prepared by UOP, LLC, December 12.

79 Elliott, D. C.; Beckman, D.; Bridgwater, A. V.; Diebold, J. P.; Gevert, S. B.; Solantausta, Y. (1991) Developments in Direct Thermochemical Liquefaction of Biomass: 1983-1990. Energy Fuels 1991, 5 (3), 399-410.

80 Elliott, D. C., Neuenschwander, G. G., Hart, T. R., Butner, R. S., Zacher, A. H., Engelhard, M. H., Young, J. S. and McCready, D. E. (2004) Chemical Processing in High-Pressure Aqueous Environments. 7. Process Development for Catalytic Gasification of Wet Biomass Feedstocks, Ind. Eng. Chem. Res. 43, 1999-2004.

81 Jaros aw Lewkowski (2001) Synthesis, chemistry and applications of 5-hydroxymethylfurfural and its derivatives ARKIVOC 17-54

82 Hanniff et al., Conversion of Biomass Carbohydrates into Hydrocarbon Products, Proceedings of IGT/CBETS Conference on Energy from Biomass and Wastes X, Washington, D.C., Apr. 7-10, 1986.

83 Diebold, J. P. and Evans, R. J., Process for producing furan from furfural aldehyde, United States Patent 4764627, August 16, 1988.

84 Lynd, L. R. (1996), Annu. Rev. Energy Environ., 21. 403-465. 85 Grethlein, A. J. and M. K. Jain (1993). Trends Biotechnol. 10. 418-423. 86 Worden, R. M., Bredwell, M. D. and Grethlein, A. J. (1997), Fuels and Chemicals from

Biomass, ACS Symposium Series No. 666. 320-336. 87 Evans, R. J. and Milne, T. A. (1987), Energy and Fuels, 1. 123-137. 88 Bridgwater, A. V. and Peacocke, G. V. C. (2000), Renewable and Sustainable Energy

Reviews. 4. 1-73. 89 Scott, D. S., Czernik, S. Piskorz, J. and Radlein, D. (1989), Energy from Biomass and Wastes

XIII, New Orleans, LA, USA, Publ by Inst of Gas Technology, Chicago, IL, USA. 90 Brown, R. C., Radlein, D., and Piskorz, J., (2001), Chemicals and Materials from Renewable

Resources: ACS Symposium Series No. 784, Washington, D.C., American Chemical Society: 123-132.

91 Prosen, E. M., Radlein, D., Piskorz, J., Scott, D. S., and Legge, R. L. (1993), Biotechnol. Bioeng., 42. 538-541.

92 Brown, R. C., Pometto, A. L., Peeples, T. L., Khiyami, M., Voss, B., Kim J. W., and Fischer, S. (2000), Proceedings of the Ninth Biennial Bioenergy Conference, Buffalo, New York.

93 Khiyami, M. A., Pometto III, A. L., Brown, R. C. , Detoxification of corn stover and corn starch pyrolysis liquors by ligninolytic enzymes of Phanerochaete chrysporium, Journal of Agricultural and Food Chemistry 53, 2969-2977, 2005.

94 MacDonald T, Yowell G, McCormack M (2001) Staff report. US ethanol industry production capacity outlook. California energy commission. Available at http://www.energy.ca.gov/reports/ 2001-08-29_600-01-017.PDF

95 Rosillo-Calle F, Cortez L (1998) Towards proalcohol II: a review of the Brazilian bioethanol programme. Biomass Bioenergy 14:115–124.

40

96 Lynd LR. Overview and evaluation of fuel ethanol from cellulosic biomass: technology, economics, the environment, and policy. Annual Reviews, Energy Environment 1996;21:403–65.

97 A. McAloon, F. Taylor, W. Yee, K. Ibsen, and R. Wooley. “Determining the Cost of Producing Ethanol from Corn Starch and Lignocellulosic Feedstocks”. National Renewable Energy Laboratory. Report, October 2000.

98 Reith JH, den Uil H, van Veen H, de Laat WTAM, Acknowledgements Niessen JJ, de Jong E, Elbersen HW, Weusthuis R, van Dijken JP and Raamsdonk L. Co-production of bio-ethanol, electricity and heat from biomass residues. In: Palz W. Spitzer J, Maniatis K, Kwant K, Helm P and Grassi A, editors. Proceedings of Twelfth European Biomass Conference. Florence Italy. ETA-Florence. 2002; 1118-1123.

99 Woods J. And Bauen A. Technology status review and carbon abatement potential of renewable transport fuels in the UK, Report prepared for Department of Trade and Industry, New and Renewable Energy Programme. Imperial College, Centre for Energy Policy and Technology. London UK. 2003; 88pp + annexes.

100 C. N. Hamelinck, G. van Hooijdonk, A. PC Faaij. “Ethanol from lignocellulosic biomass: techno-economic performance in short-, middle-, and long-term”. Biomass and Bioenergy 28, (2005) 384-410.

101 Wyman CE, Bain RL, Hinman ND, Stevens DJ. Ethanol and Methanol from Cellulosic Biomass. In: Johansson TB, Kelly H, Reddy AKN, Williams RH, Burnham L, Eds. Renewable energy, Sources for fuels and electricity. Washington DC, USA: Island Press; (1993) p. 865.

102 R. Wooley, M. Ruth, D. Glassner, J. Sheehan. Process design and costing of bioethanol technology: a tool for determining the status and direction of research and development. Biotechnology Progress. (1999) 15 p. 794.

103 L.R. Lynd, R. T. Elander, C. E. Wyman. Likely features and costs of mature biomass ethanol technology. Applied Biochemistry and Biotechnology (1996) 57/58 741.

104 Developments in Thermochemical Biomass Conversion; A. V. Bridgwater, and D. G. B. Boocock, , Eds.; Blackie Academic & Professional, London (1997) pp. 391.

105 E. D. Larson, R. E. Katofsky. Production of Methanol and Hydrogen from Biomass. The Center for Energy and Environmental Studies, Princeton University, PU/CEEs Report no. 217, July.

106 M. K. Mann, (1995a). “Technical and economic assessment of producing hydrogen by reforming syngas from the Battelle indirectly heated biomass gasifier.” National Renewable Energy Lab, NREL/TP-431-8143.

107 R. E. Katofsky, The Production of Fluid Fuels from Biomass, Center for Energy and Environmental Studies, Princeton University, Princeton, (1993).

108 H. Komiyama, T. Mitsumori, K. Yamaji, K. Yamada, Assessment of energy systems by using biomass plantation, Fuel 80 (2001) 707.

109 C. N. Hamelinck, and A. Faaij. “Future prospects for production of methanol and hydrogen from biomass”. Journal of Power Sources 111 1 (2002).

110 R. E. Katofsky, The Production of Fluid Fuels from Biomass, Center for Energy and Environmental Studies, Princeton University, Princeton (1993).

111 R. H. Williams, E.D. Larson, R.E. Katofsky, J. Chen, Methanol and hydrogen from biomass for transportation, with comparisons to methanol and hydrogen from natural gas and coal,

41

PU/CEES Report 292, Center for Energy and Environmental Studies, Princeton University, Princeton, NJ (1995) p. 47.

112 “Mixed Alcohols from Syngas: State of Technology” Nexant, NREL Contract ACO-5-44027, May 2005.

113 Nirula, S. C. (1994). "Dow/Union Carbide Process for Mixed Alcohols from Syngas." PEP Review no. 85-1-4, SRI International, Meno Park, CA.

114 Olayan, H. B. M. (1987). "Selection of technology for synthesis gas based products in Saudi Arabia." Energy Progress, 7(1), 9-17.

115 Courty, P., Chaumette, P., Raimbault, C., and Travers, P. (1990). "Production of methanol-higher alcohol mixtures from natural gas via syngas chemistry." Revue de l'Institut Francais du Petrole, 45(4), 561-78.

116 S. Phillips, A. Aden, J. Jechura, D. Dayton, and T. Eggeman. “Thermochemical Ethanol via Indirect Gasification and Mixed Alcohol Synthesis of Lignocellulosic Biomass”. NREL/TP-510-41168 April 2007.

117 “Economics of MTBE via Mixed Alcohol Synthesis”Alternative Fuels and Chemicals from Synthesis Gas, cooperative agreement DE-FC22-95PC93052. 1996.

118 Fluor 2003 Gasification Technologies Conference paper, ibid. 119 “Optimization of Electricity-Methanol Coproduction. Configurations of Integrated-

Gasification-Combined-Cycle/Once-Through.” ChemSystems. EPRI/GS-6869, 1990. 120 Salmon, R. "Economics of Methanol Production from Coal and Natural Gas." ORNL-6091, Oak Ridge National Laboratory. 1986.

121 Courty, P., Arlie, J. P., Convers, A., Mikitenko, P., and Sugier, A. "C1-C6 Alcohols from Syngas." Hydrocarbon Processing, 63(11), 105-108. 1984.

122 Ricci, R., Paggini, A., Fattore, V., Ancillotti, F., and Sposini, M. "Production of methanol and higher alcohols from synthesis gas." Chemia Stosowana, 28(1), 155-68. 1984.

123 El Sawy, A. H. "Evaluation of Mixed Alcohol Production Processes and Catalysts." NTIS. DE90010325. SAND89-7151, Mitre Corporation. 1990.

124 Bechtel Corporation. "Task 4.2 Commercial Applications - Economics of MTBE via Mixed Alcohol." Subcontract No. PT5781-B, Prepared for Air Products and Chemicals. 1998.

125 Lutz, B. (2001). "New age gas-to-liquids processing." Hydrocarbon Engineering 6(11): 23-26, 28. 126 Schulz H. Short history and present trends of FT synthesis. Applied Catalysis A: General

1999;186:1–16.127 So, K.S., Brown, R. C. (1999) Economic analysis of selected lignocellulose to ethanol

conversion technologies, Applied Biochemistry and Biotechnology 77-79, 633-640. 128 Brown, R. C. (2007) Hybrid thermochemical/biological processing of biomass, Applied

Biochemistry and Biotechnology 137-140, 947 – 956.

49

Review

Comparative economics of biorefi neries based on the biochemical and thermochemical platformsMark M. Wright and Robert C. Brown, Center for Sustainable Environmental Technologies, Iowa State University

Received 17 May 2007; revised version received 11 June 2007; accepted 11 June 2007

Published online in Wiley InterScience (www.interscience.wiley.com); DOI: 10.1002/bbb.8;

Biofuels, Bioprod. Bioref. 1:49–56 (2007)

Abstract: A variety of biochemical and thermochemical technologies have been proposed for the production of

biofuels. Meaningful economic comparisons require that they be evaluated on the same bases in terms of technol-

ogy maturity, plant capacity, the energy content of the fuel, feedstock costs, method of calculating capital charges,

and year in which the analysis is assumed. Such an analysis for fi rst-generation biomass-to-biofuels plants reveals

that capital costs will be comparable for advanced biochemical and thermochemical biorefi neries, costing four to

fi ve times as much as comparably sized grain ethanol plants. The cost of advanced biofuels, however, will be similar

to that of grain ethanol as corn prices exceed $3.00 per bushel. © 2007 Society of Chemical Industry and John

Wiley & Sons, Ltd

Keywords: biochemical platform; thermochemical platform; biofuel; economics

Introduction

The commercial success of the grain ethanol industry has increased interest in processes that convert fi brous biomass (lignocellulose) into biofuels. Much of this

attention has been directed towards enzymatic hydrolysis of cellulose to simple sugars and its subsequent fermentation to ethanol, the so-called biochemical platform. However, the recent announcement by the Department of Energy that three of the six biorefi nery projects selected for federal

funding include gasifi cation technologies has increased interest in the thermochemical platform.1 Th is approach thermolytically transforms biomass into gaseous or liquid intermediate chemicals that can be upgraded to transporta-tion fuels or commodity chemicals. Th e thermochemical platform is related to the commercially successful proc-esses used by the petroleum and petrochemical industries to transform fossil fuels into fuels and chemicals, although biomass and coal have very diff erent physical and chemical properties and their processing will diff er in some respects.

Correspondence to: Robert C. Brown, Center for Sustainable Environmental Technologies, 286 Metals Development,

Iowa State University, Ames, IA-50010, USA. E-mail: rcbrown@iastate.edu

© 2007 Society of Chemical Industry and John Wiley & Sons, Ltd

50

MM Wright, RC Brown Review: Comparative economics of biorefineries

Opinions vary widely on which of these platforms is most likely to prevail in the emergence of cellulosic biofuels. Although a number of techno-economic studies of various manifestations of these two platforms have appeared in the literature, direct comparisons of them on the same bases are diffi cult to fi nd.

Th is paper compares the capital costs and operating costs of the current generation of starch-based ethanol plants with those of fi rst-generation lignocellulose-to-biofuels plants. Th ese lignocellulosic biofuels platforms include enzymatic hydrolysis of cellulose to ethanol (the biochemical platform) and three manifestations of the thermochemical platform in which biomass is gasifi ed and upgraded to hydrogen, meth-anol, or Fischer–Tropsch (F-T) liquids.

Background

Current ethanol production is based on either sugar or starch crops, with the former dominating in Brazil and the latter in the USA, the two largest commercial ethanol producers in the world. Th e present study focuses on dry milling of corn (a starch crop) because of its relevance to US markets. Although wet milling has some effi ciency and production cost advantages, dry milling has dominated the recent grain ethanol industry because of its lower require-ments for capital and labor.

Dry milling consists of four major operations: grinding to make the starch accessible to enzymes; cooking with enzymes to hydrolyze the starch to sugars; fermentation of sugars by yeast to produce ethanol, and distillation to produce neat ethanol. Th e byproduct of this process is distillers’ dried grains, a fi ber and protein-rich material that is used as livestock feed.

Techno-economic analyses of dry grind ethanol in the USA and Europe estimate production costs to be between $0.80 and $1.36 per gallon of ethanol depending upon assumptions, especially feedstock costs.2–4 Th e present study employs, for the grain ethanol base case, the oft en cited 2000 report by McAloon and colleagues at the National Renewable Energy Laboratory.5 Th is report estimates that a dry grind corn processing plant producing 25 million gallons of ethanol per year would cost $27.9 million (1999, USA). Th e operating cost of this plant was estimated to be $0.88 per gallon of ethanol based on feedstock cost of $1.94 per bushel of corn grain.

Th e advanced biochemical platform employs cellulose (and hemicellulose) from plant fi bers instead of starch in the production of ethanol. Th e process of biologically converting structural carbohydrate into ethanol consists of four major operations: pretreatment of plant fi bers to make cellulose accessible to enzymes (which oft en chemically hydrolyzes hemicellulose to simple sugars); treatment with enzymes to hydrolyze cellulose to glucose; fermentation of the simple sugars to produce ethanol; and distillation to produce neat ethanol. Lignin is a byproduct of this process, which can be used as boiler fuel.

Based on the current state of technology, capital costs for biochemical ethanol from cellulose are estimated to be between $4.03 and $5.60 per gallon of ethanol annual capacity.6–9 Operating costs are estimated to be between $1.34 and $1.69 depending upon the assumptions made about feedstock costs, enzyme costs, and the kind of pretreatment to be employed.6–9 Th ese cost estimates refl ect early implementations of cellulose-based technologies. Projected capital cost for future plants employing antici-pated improvements in conversion technologies are esti-mated to be $3.33–4.44 per gallon ethanol annual capacity with operating costs dropping to $0.40 and $0.89 per gallon of ethanol.10 Th e present study uses the recent analysis of Hamelinck et al.,10 which builds upon earlier studies and employs more recent information on the unit operations employed for fi rst-generation technology. Th e process is based on dilute acid pretreatment and enzymatic hydrolysis. Th e total capital investment was calculated to be $294 million, while the operating costs were an estimated $1.51 per gallon of ethanol.

Unlike the biochemical platform, for which the fuel product is defi ned (ethanol) and there is reasonable consensus as to the unit operations to be employed, there is greater diver-sity of opinion on how the thermochemical platform should be confi gured. Th ermolytic processing of biomass can generally be categorized as either gasifi cation (to produce a gaseous product) or fast pyrolysis (to produce primarily liquid product). A variety of catalytic11 or even biocatalytic12 schemes have been proposed to upgrade the thermolytic products into alcohols, ethers, esters, or hydrocarbons. Th e present analysis focuses on gasifi cation followed by catalytic synthesis to three fuels: hydrogen, methanol, and F-T liquids.

© 2007 Society of Chemical Industry and John Wiley & Sons, Ltd | Biofuels, Bioprod. Bioref. 1:49–56 (2007); DOI: 10.1002/bbb

© 2007 Society of Chemical Industry and John Wiley & Sons, Ltd | Biofuels, Bioprod. Bioref. 1:49–56 (2007); DOI: 10.1002/bbb 51

Review: Comparative economics of biorefineries MM Wright, RC Brown

Hydrogen can be manufactured from syngas via the water–gas shift reaction:

CO + H2O → CO2 + H2

Th is reaction requires the mixing of steam with syngas since biomass gasifi cation rarely releases suffi cient water vapor for this purpose. Although hydrogen might be one of the most cost-eff ective clean-burning biofuels to produce, the physical characteristics of hydrogen present challenges in its use as transportation fuel.13

Methanol is commercially manufactured from syngas using a copper-based catalyst via the reaction:14

CO + 2H2 → CH3OH

As a transportation fuel, it has many of the same advan-tages and disadvantages as ethanol.15 However, methanol is considerably more toxic than ethanol. Recent rulings by the US Environmental Protection Agency (EPA) are likely to ban the closely related and similarly toxic methyl tertiary butyl ether (MTBE) as a fuel additive because of concerns about groundwater contamination.16

Fischer–Tropsch liquids are synthetic hydrocarbon fuels produced from syngas by the action of metal catalysts at elevated pressures. Th e primary products of F-T synthesis are a mixture of light hydrocarbon gases, paraffi n waxes, and alcohols according to the generalized reaction:17

CO + 2H2 → – CH2 – + H2O

Th e optimal H2:CO ratio of 2:1 is achieved through the water–gas shift reaction ahead of the synthesis reactor. Depending on the types and quantities of F-T products desired, either low (200–240°C) or high (300–350°C) temperature synthesis is used with either an iron (Fe) or cobalt catalyst (Co). Additional processing of the F-T prod-ucts yields diesel fuel or gasoline.

Th e fi ve major unit operations of the thermochemical platform based on gasifi cation include fuel preparation, gasifi cation, gas clean-up, catalytic processing to the desired fuels, and separations. Fuel preparation is typically size reduction and drying to levels consistent with the gasifi -cation technology employed. Although a large variety of biomass gasifi cation technologies can be envisioned,18 the

thermochemical base cases included in this analysis assume oxygen-blown, high-temperature gasifi cation with the F-T process also employing pressurized gasifi cation. Gas clean-up includes removal of particulate matter and trace contam-inants including sulfur, chlorine, and ammonia. Separations are designed to yield pure fuel. Catalytic processing may require multiple catalysts operating at diff erent conditions of temperature and pressure. Th e water–gas shift reaction is common to the production of most synfuels, including hydrogen, methanol, and F-T liquids. Methanol synthesis is optimal at a syngas H2:CO ratio of 3:1 while F-T synthesis favors a H2:CO ratio of 2.15.19 Of course, hydrogen produc-tion involves the complete reaction of CO and steam to form hydrogen fuel. Separations are an integral part of thermochemical processing but, unlike the aqueous phase processes of the biochemical platform, thermochemical processes employ vapor-phase reactions that do not require energy-intensive distillations to remove water from the fuel product. Although some would argue that the ther-mochemical platform is technically mature because of its commercial implementation using coal as feedstock as early as the 1930s, this presumes a close similarity between coal and biomass feedstocks. In fact, they diff er considerably in volatile content, char reactivity, and mineral content. Th us, like the biochemical platform, technical improvements are anticipated through research and development, although no attempt is made to anticipate these improvements in the present analyses.

Th e literature includes several studies on the cost of hydrogen from biomass based on the present state of the art in gasifi cation and synfuels catalysts. Capital costs range from $0.65 to $1.33 per gallon of liquid hydrogen capacity20, 21 depending upon the type and size of gasifi er plant. Operating costs range from $0.31 to $0.44 per gallon of liquid hydrogen produced22 depending upon the cost of biomass and the kinds of processes employed.23 Th e study by Hamelinck and Faaij24 is one of the most recent studies (2002) on biomass-to-hydrogen and includes data in suffi -cient detail to serve as the base case for the present study. For a gasifi cation plant producing 220 million gallons of liquid hydrogen fuel per year, the estimated capital cost is $206 million with operating costs of $0.24 per gallon of liquid hydrogen produced.

52 © 2007 Society of Chemical Industry and John Wiley & Sons, Ltd | Biofuels, Bioprod. Bioref. 1:49–56 (2007); DOI: 10.1002/bbb

MM Wright, RC Brown Review: Comparative economics of biorefineries

Early techno-economic studies of methanol plants report that production costs range from $0.91 to $1.11 per gallon of methanol.22, 25 Th e techno-economic analysis of Hamelinck and Faaij was chosen for this study because it analyzes both hydrogen and methanol plants. Based on averages from their analysis, a gasifi cation plant producing 87 million gallons of methanol per year would require a capital investment of $276 million. Th e production cost of methanol was $0.62 per gallon of methanol.

Previous studies of gasifi cation plants producing F-T diesel suggest that costs range from $1.1 to $4.1 per gallon of F-T diesel.26, 27 Among the most recent and complete techno-economic analyses of a F-T diesel plant is that by Tijmensen et al.28 Th is analysis employs the oxygen-blown, pressurized gasifi er of the Institute of Gas Technology confi gured to achieve the preferred H2 to CO ratio. A reformer and water shift reactor are not necessary for this confi guration. Th e product selectivity for hydrocarbon chains of fi ve carbons or longer is 73.7 to 91.9%. Capital cost for a plant producing about 35 million gallons of F-T diesel per year is estimated

to be about $341 million with operating costs of about $2.37 per gallon of F-T diesel.

Table 1 summarizes the data on plant capacity, capital cost, fuel effi ciency, and heating value of the fuel produced for each of the fi ve biofuel plants analyzed in this study. Fuel effi ciency is defi ned as the fraction of total energy inputs that appear as chemical energy of the product fuel, which in this analysis is neat ethanol, hydrogen, meth-anol, or F-T liquids. Th e energy inputs can come from either fossil energy or renewable energy. For example, the production of grain ethanol has both biomass energy inputs (grain) and fossil energy inputs (natural gas or coal) while all four cellulosic biofuels processes obtain their energy requirements exclusively from biomass. Energy outputs other than biofuels, such as process heat, electricity, and distillers’ dried grains, do not fi gure in fuel effi ciency calcu-lations but count towards credits in calculating net operating costs.

Table 2 breaks down the operating costs for each plant. Clearly, direct comparisons among these fi ve plants are

Table 1. Capital costs of reference plants.

Plant typePlant capacity

(MMGPY)*Capital cost

($ million)Fuel effi ciency

(%)** Basis yearFuel heating value (MJ/L)

Source reference

Grain ethanol 25 27.9 35 1999 21 –

Cellulosic ethanol 50 294 35 2005 21 10

Methanol 87 254 45 2002 16 24

Hydrogen 182 244 50 2002 8 4

Fischer–Tropsch 35 341 46 2002 36 28

* Millions of gallons per year; all fuels assumed to be liquefi ed.

** Defi ned as the fraction of energy inputs into the plant that appears as chemical energy of the product fuel.

Table 2. Operating cost components of reference plants.

Plant Type

Biomassfeedstock($ million)

Operation and management

($ million)Credits

($ million)Capital charges

($ million)Total

($ million)Grain Ethanol 17.0 10.5 $7.1 2.8 22.0

Cellulosic Ethanol 35.4 11.1 * 29.4 76.0

Methanol 24.8 10.2 $8.68 30.2 56.5

Hydrogen 24.7 9.76 $9.88 29.0 53.6

Fischer-Tropsch 29.2 14.6 * 43.8 87.5

* Byproducts or waste energy are used for process heat within the plant.

© 2007 Society of Chemical Industry and John Wiley & Sons, Ltd | Biofuels, Bioprod. Bioref. 1:49–56 (2007); DOI: 10.1002/bbb 53

Review: Comparative economics of biorefineries MM Wright, RC Brown

diffi cult without placing them on a common basis, as described in the methodology section.

Methodology

Th e approach to this study was to place existing techno-economic analyses in the published literature on the same bases. Th e fi rst step was to identify studies that assumed the deployment of fi rst-generation technology for the produc-tion of biofuels from lignocellulosic biomass. Although some studies anticipate technology improvements and estimate the costs for mature biofuels plants, the present study is limited to state-of-the-art technology for which more reli-able data is available. Th e data for these baseline cases were adjusted to account for diff erences in plant capacity, the gasoline equivalency of the fuel, feedstock costs, method of calculating capital charges, and the year for which costs were estimated. Readers desiring further details on the techno-economic analyses employed in this study should refer to the original publications.

Since economies of scale strongly infl uence the cost of production, capital costs in the original studies were adjusted to a common plant size of 150 million gallons per year gasoline equivalent. Th is size was selected as represent-ative of expected early generation cellulosic biofuel plants, but no attempt was made to optimize plant size (which will be diff erent for each of the technologies employed). Capital costs were scaled to plant size using a simple power law commonly employed to account for economies of scale.29 A scaling exponent of 0.63 was assumed5 for the grain ethanol and cellulosic ethanol (biochemical) platforms while a scaling exponent of 0.7 was assumed6 for the thermo-chemical platforms. Th ese exponents were selected based on recommendations from the referenced studies and refl ect the current understanding of the scaling behavior of these respective technologies. Following usage in the petroleum industry, capital costs are expressed as dollars ‘per barrel per day’ (pbpd) of production capacity. Th e capacity factor of the plant is assumed to be 0.9.

Plant capacity is customarily reported as the volume of fuel produced although the volumetric energy density (MJ/L) can vary considerably among diff erent kinds of fuels. For example, ethanol, the most prominent biofuel manufac-tured today, has only 2/3 the enthalpy of an equal volume of

gasoline. Th us, production capacity and production cost are reported in this paper on the basis of the gasoline equiva-lence of the fuel produced.

Operating costs are broken down into categories of biomass feedstock, operation and management, byproduct credits, and capital charges. Biomass costs are propor-tional to plant capacity and process effi ciency in converting biomass into fuels. Cellulosic feedstock costs are assumed to be $50 Mg–1. Corn is priced at $2.12 per bushel, the price prevailing in the basis year (2005). Operation and manage-ment costs include plant and management labor, materials and supplies to operate the plant, and utilities. Credits are given in some instances for byproducts. For example, grain ethanol produces distillers’ dried grains, which can be sold as cattle feed. Gasifi cation generates high temperature heat that can be used for electric power generation. A realistic analysis of operating costs must include the cost of capital although this calculation can be complicated by the fact that actual projects are usually fi nanced by a combination of debt and investor capital. To simplify the evaluation, this study assumes 100% debt fi nancing over 20 years at an annual interest rate of 8%. Infl ation can make it diffi cult to compare studies performed in diff erent years. Accordingly, all costs were adjusted to 2005 dollars.

Results

Th e capital costs and operating costs for the various biochemical and thermochemical biofuels plants on a common basis are compared in Table 3. It is clear that advanced biofuels will come at very high capital cost–more than fi ve times that of comparably sized starch ethanol plants–based on the current state of technology. In terms of least capital cost, the order of preference for cellulosic biofuels is thermochemical hydrogen, methanol, lignocel-lulosic ethanol, and F-T diesel. Th e diff erence in capital costs among the cellulosic biofuel options is signifi cant: F-T diesel requires almost 50% greater investment than thermochemical hydrogen. Th is diff erence refl ects the addi-tional unit operations required to convert syngas into F-T diesel. Th e two most fungible fuels among the advanced biofuels, lignocellulosic ethanol and F-T diesel, are the most capital-intensive processes. At $76,000 pbpd and $86,000 pbpd, respectively, their capital costs are essentially the

54 © 2007 Society of Chemical Industry and John Wiley & Sons, Ltd | Biofuels, Bioprod. Bioref. 1:49–56 (2007); DOI: 10.1002/bbb

MM Wright, RC Brown Review: Comparative economics of biorefineries

same within the uncertainty of the analysis (+/ – 30%). In comparison, the capital cost for a grain to ethanol plant is only $13,000 pbpd.

Th e rank ordering in terms of operating costs is the same as for capital costs, with thermochemical hydrogen being the least expensive followed by methanol, lignocellulosic ethanol, and F-T diesel. Operating costs range from a low of $1.05 per gallon of gasoline equivalent for hydrogen to $1.80 per gallon of gasoline equivalent for F-T diesel. In compar-ison, grain ethanol for this size of plant could be produced for $1.22 per gallon of gasoline equivalent. Th is is cheaper than all the cellulosic biofuels except hydrogen.

Figure 1 helps one to understand the diff erences in oper-ating costs for the various biofuels plants. Much of the advantage of thermochemical hydrogen comes from its relatively low biomass costs, which arises from its high fuel effi ciency. As shown in Table 1, thermochemical hydrogen has a fuel effi ciency of 50% compared to about 45% for the other two thermochemical technologies. Whereas hydrogen production from syngas requires only water–gas shift reac-tion and gas separation, methanol production requires an additional catalytic step, and F-T diesel requires at least two additional catalytic steps, each with attendant loses in effi -ciency. At 35%, lignocellulosic ethanol is even less effi cient,

Table 3. Capital cost and operating costs for 150 MMGPY gasoline equivalent plants (2005 dollars).

FuelTotal capital cost

($ millions)Capital cost per unit production (pbpd)*

Operating cost ($ per gallon)**

Grain ethanol 111 13,000 1.22

Cellulosic ethanol 756 76,000 1.76

Methanol 606 66,000 1.28

Hydrogen 543 59,000 1.05

Fischer-Tropsch 854 86,000 1.80

* Per barrel per day gasoline equivalent.

** Gallons gasoline equivalent.

Figure 1. Operating costs for 150 MMGPY of gasoline equivalent.

-$1.00

-$0.50

$0.00

$0.50

$1.00

$1.50

$2.00

GrainEthanol

CellulosicEthanol

Methanol Hydrogen Fischer-Tropsch

Op

erat

ing

Co

st (

$/g

allo

n o

f g

aso

line

equ

ival

ent)

Credits

Operation andManagementBiomass Costs

Capital Costs

© 2007 Society of Chemical Industry and John Wiley & Sons, Ltd | Biofuels, Bioprod. Bioref. 1:49–56 (2007); DOI: 10.1002/bbb 55

Review: Comparative economics of biorefineries MM Wright, RC Brown

which arises from the inability to convert the non-carbohy-drate fraction of lignocellulose into biofuel.

Grain ethanol has the highest biomass costs among the fi ve technologies evaluated. Th is refl ects a combination of relatively low fuel effi ciency (about one-third of the corn grain ends up as the byproduct distillers’ dried grains) and high fuel cost (corn grain at $2.12 per bushel is almost 75% more than lignocellulosic biomass on a dry weight basis). However, as Fig. 1 illustrates, the byproduct from a dry mill corn ethanol plant (distillers’ dried grains) yields a produc-tion credit almost three times greater than that achieved by any of the other processes. If the expanding grain ethanol industry produces an oversupply of distillers’ dried grains (assumed to be worth $99 Mg–1 in the present study), the attractive production cost for corn ethanol could evaporate.

Another scenario that could diminish the attractive production cost of grain ethanol is already developing. Th e present analysis was based on the 2005 price for corn grain, which was only $2.12 per bushel. Substituting $3.00 per bushel, which is more typical of the selling price in late 2006, increases production cost of grain ethanol to $1.74 per gallon of gasoline equivalent, which is comparable to the price for cellulosic ethanol and F-T diesel.

Conclusions

Th e rapidly expanding renewable fuels industry will soon have to turn to technologies that convert lignocellulosic biomass into biofuels. Although much of the attention on advanced biofuels has focused on lignocellulosic ethanol, thermolytic processes have also been proposed for produc-tion of renewable fuels. Meaningful comparison of advanced fuels technologies with the current grain ethanol process requires that techno-economic analysis be on the same basis.

Such an adjustment to techno-economic analyses has been done for 150 million gallon (gasoline equivalent), fi rst-generation lignocellulosic biofuels plants based on biochem-ical production of ethanol and thermochemical production of hydrogen, methanol, and F-T diesel. Th is analysis reveals that advanced biofuels will come at very high capital cost – more than fi ve times that of comparably sized starch ethanol plants. Th us, raising the $0.5 billion to almost $1 billion in capital for a cellulosic biofuels plant will be much more

diffi cult than has proved to be the case for grain ethanol plants. Th e smaller number of unit operations associated with thermochemical hydrogen makes it the least capital intensive of the four advanced biofuels options evaluated, including lignocellulosic ethanol. However, the larger prob-lems of hydrogen storage and distribution infrastructure compared to other fuels makes it less likely to be adopted within the 10-year time frame envisioned for signifi cant expansion of biofuels in the USA. Th e capital costs for ligno-cellulosic ethanol and F-T diesel, probably the most fungible fuels among the four advanced biofuels options considered, are essentially the same within the uncertainty of the anal-ysis, costing about $80,000 pbpd.

Th ermochemical hydrogen also has the lowest production cost, at about $1.05 per gallon of gasoline equivalent, of the four advanced biofuels options for a plant of 150 million gallons gasoline equivalent annual capacity. It even bests grain ethanol, which would cost $1.27 per gallon of gaso-line equivalent. Lignocellulosic ethanol and F-T diesel have almost the same production costs, at about $1.78 per gallon of gasoline equivalent. If corn prices increase from $2.12 per bushel that prevailed in the basis year of this analysis (2005) to $3.00 per bushel found in late 2006 and expected in future years because of increased corn demand, then biofuels from lignocellulosic biomass will have comparable costs of production. Future improvements in the biochemical and thermochemical platforms for processing lignocellulose have the potential to further reduce the costs of biofuels.

From these comparisons it is concluded that, for the current state of the technology, neither the biochemical nor thermochemical platforms have clear advantages in capital costs or operating costs for production of advanced biofuels. Both technologies have opportunities to compete against grain ethanol as corn prices continue to rise, especially if the high capital costs of advanced biofuels plants can be dramatically reduced.

Acknowledgements

Th is study was supported by the Iowa Energy Center under Contract No. 0606 and the USDA under Contract No. 5819356654. We would like to acknowledge the encourage-ment and support of Norm Olson of the Iowa Energy Center and Akwasi Boateng of the Eastern Regional Research Center, ARS, USDA.

56 © 2007 Society of Chemical Industry and John Wiley & Sons, Ltd | Biofuels, Bioprod. Bioref. 1:49–56 (2007); DOI: 10.1002/bbb

MM Wright, RC Brown Review: Comparative economics of biorefineries

References 1. Calamur K, Analysis: Biorefi nery boost for ethanol, United Press

International, February 28 (2007).

2. Reith JH, den Uil H, van Veen H and de Laat WTAM, Co-production of

bio-ethanol, electricity and heat from biomass residues, in Proceedings

of Twelfth European Biomass Conference, Florence, Italy, ed. by Palz W,

Spitzer J, Maniatis K, Kwant K, Helm P and Grassi A. ETA, Florence,

pp. 1118–1123 (2002).

3. Woods J and Bauen A, Technology status review and carbon abatement

potential of renewable transport fuels in the UK. Report prepared

for Department of Trade and Industry, New and Renewable Energy

Programme. Imperial College, Centre for Energy Policy and Technology,

London.

4. Wooley R, Ruth M, Sheehan J, Ibsen K, Majdeski H and Galvez A,

Lignocellulosic biomass to ethanol—Process design and economics

utilizing co-current dilute acid prehydrolysis and enzymatic hyrolysis—

Current and futuristic scenarios. Report No. TP-580-26157, National

Reneawable Energy Laboratory, Golden Colorade, USA (1999).

5. McAloon A, Taylor F, Yee W, Ibsen K and Wooley R, Determining the cost

of producing ethanol from corn starch and lignocellulosic feedstocks.

National Renewable Energy Laboratory, Report, October (2000).

6. Wyman CE, Bain RL, Hinman ND and Stevens DJ, Ethanol and methanol

from cellulosic biomass, in Renewable Energy, Sources for Fuels and

Electricity, ed. by Johansson TB, Kelly H, Reddy AKN, Williams RH and

Burnham L. Island Press, Washington DC (1993).

7. Wooley R, Ruth M, Glassner D and Sheehan J, Process design and

costing of bioethanol technology: a tool for determining the status and

direction of research and development. Biotechnol Progress 15:794

(1999).

8. Lynd LR, Elander RT and Wyman CE, Likely features and costs of mature

biomass ethanol technology. Appl Biochem Biotechnol 57/58:741 (1996).

9. Aden A, Ruth M, Ibsen K, Jechura J, Neeves K, et al., Lignocellulosic

biomass to ethanol process design and economics utilizing co-current

dilute acid prehydrolysis and enzymatic hydrolysis for corn stover.

National Renewable Energy Laboratory, Report, June (2002).

10. Hamelinck CN, van Hooijdonk G and Faaij AOC, Ethanol from

lignocellulosic biomass: techno-economic performance in short-,

middle-, and long-term. Biomass Bioenergy 28:384 (2005).

11. Bridgwater AV and Boocock DGB (eds), Developments in

Thermochemical Biomass Conversion. Blackie Academic & Professional,

London (1997).

12. Brown RC, Biomass refi neries based on hybrid thermochemical/

biological processing – an overview, in Biorefi neries, Biobased Industrial

Processes and Products, ed by Kamm B, Gruber PR and Kamm M. Wiley-

VCH Verlag, Weinheim, Germany (2005).

13. Romm JJ, The Hype About Hydrogen: Fact and Fiction in the Race to

Save the Climate, Island Press, Washington, DC (2004).

14. Yang H and Liao P, Preparation and activity of Cu/ZnO-CNTs nano-

catalyst on steam reforming of methanol. Appl Catal A: Gen 317(2): 226

(2007).

15. Gandhidasan P, Ertas A and Anderson EE, Review of methanol and

compressed natural gas (CNG) as alternative for transportation fuels.

J Energy Resource Technol, Trans ASME 113(2):101 (1991).

16. Davenport RE, Gubler R and Yoneyama M, Chemical Economics

Handbook Marketing Research Report - Ethyl Alcohol, SRI International,

May (2002).

17. Dry ME. Sasol Fischer-Tropsch processes. Appl Ind Catal 2:167–213

(1983).

18. Brown RC, Biomass energy conversion, Section 24.2 Power generation,

in CRC Handbook of Energy Conservation and Renewable Energy, ed by

Kreith F and Goswami Y. CRC Press (2006).

19. Spatch PL and Dayton DC, Preliminary screening – technical and

economic assessment of synthetic gas to fuels and chemicals with

emphasis on the potential for biomass-derived syngas. NREL Technical

Report (2003).

20. Larson ED and Katofsky RE, Production of methanol and hydrogen from

biomass. The Center for Energy and Environmental Studies, Princeton

University, PU/CEEs Report No. 217, July (1992).

21 Mann MK, Technical and economic assessment of producing hydrogen

by reforming syngas from the Battelle indirectly heated biomass gasifi er.

National Renewable Energy Lab, NREL/TP-431-8143 (1995).

22. Katofsky RE, The production of fl uid fuels from biomass. Center for

Energy and Environmental Studies, Princeton University, Princeton NJ

(1993).

23. Komiyama H, Mitsumori T, Yamaji K and Yamada K, Assessment of

energy systems by using biomass plantation. Fuel 80: 707 (2001).

24. Hamelinck CN and Faaij A, Future prospects for production of methanol

and hydrogen from biomass. J Power Sources 111: 1 (2002).

25. Williams RH, Larson ED, Katofsky RE and Chen J, Methanol and

hydrogen from biomass for transportation, with comparisons to methanol

and hydrogen from natural gas and coal. PU/CEES Report 292, Center

for Energy and Environmental Studies, Princeton University, Princeton,

NJ (1995).

26. MITRE Corporation, Techno-economic Assessment of Biomass

Gasifi cation Technologies for Fuels and Power (1996).

27. Netherlands Agency for Energy and the Environment (2000). Technical

and Economic Data Biomass-Bases Energy Conversion: Systems for the

Production of Gaseous and/or Liquid Energy Carriers. Technical Report

No. GAVE-9915.

28. Tijmensen JAM, Faaij APC and Hamelinck CN, Exploration of the

possibilities for production of Fischer Tropsch liquids and power via

biomass gasifi cation. Biomass Bioenergy 23: 129 (2002).

29. Brown RC, Biorenewable Resources: Engineering new Products from

Agriculture. Iowa State Press (2003).

191

Review

Establishing the optimal sizes of different kinds of biorefi neries Mark Wright and Robert C. Brown, Iowa State University, Ames, USA

Received August 3, 2007; revised version received August 19, 2007; accepted August 20, 2007

Published online October 9, 2007 in Wiley InterScience (www.interscience.wiley.com); DOI: 10.1002/bbb.25;

Biofuels, Bioprod. Bioref. 1:191–200 (2007)

Abstract: This paper explores the factors that infl uence the optimal size of biorefi neries and the resulting unit cost

of biofuels produced by them. Technologies examined include dry grind corn to ethanol, lignocellulosic ethanol via

enzymatic hydrolysis, gasifi cation and upgrading to hydrogen, methanol, and Fischer Tropsch liquids, gasifi cation

of lignocellulosic biomass to mixed alcohols, and fast pyrolysis of lignocellulosic biomass to bio-oil. On the basis of

gallons of gasoline equivalent (gge) capacity, optimally sized gasifi cation-to-biofuels plants were found to be

50–100% larger than biochemical cellulosic ethanol plants. Biorefi neries converting lignocellulosic biomass into

transportation fuels were found to be optimally sized in the range of 240–486 million gge per year compared to

79 million gge per year for a grain ethanol plant. Among the biofuel options, ethanol, whether produced biochemi-

cally or thermochemically, is the most expensive to produce. Lignocellulosic biorefi neries will require 4.7–7.8 million

tons of biomass annually compared to 1.2 million tons of corn grain for a grain ethanol plant. Factors that could

reduce the optimal size of lignocellulosic biorefi neries are discussed. © 2007 Society of Chemical Industry and John

Wiley & Sons, Ltd

Keywords: biorefi nery; optimal size; unit cost; gallons of gasoline equivalent

Introduction

The petroleum-based motor fuels industry is charac-terized by giant refi neries, processing petroleum at rates equivalent to ten gigawatts of power or more

(140 000 barrels of petroleum per day). Th is situation exists because operating costs are driven by economies of scale, which causes operating costs to increase more slowly as plant size gets larger. Th us, unit costs for transportation fuels and

commodity chemicals derived from fossil fuels are expected to decrease monotonically with increasing plant size.

Th e situation is dramatically diff erent for biorefi neries where biomass feedstocks are obtained from a multiplicity of ‘farm gates’ as opposed to a single ‘mine mouth’ or ‘well head’ of a refi nery based on fossil fuels. Th ese farm gates are widely distributed geographically, resulting in transpor-tation costs to the plant that strongly depend upon the size of the processing plant. Furthermore, transport of solid,

*Correspondence to: Robert C. Brown, Iowa State University, Center for Sustainable Environmental Technologies,

285 Metals Development Bldg, Ames, IA 50011-3020, USA. E-mail: rcbrown@iastate.edu

© 2007 Society of Chemical Industry and John Wiley & Sons, Ltd

192 © 2007 Society of Chemical Industry and John Wiley & Sons, Ltd | Biofuels, Bioprod. Bioref. 1:191–200 (2007); DOI: 10.1002/bbb

M Wright, RC Brown Review: Optimal sizes of different kinds of biorefineries

low density biomass is more labor intensive and expensive than the movement of gaseous and liquid fuels, like natural gas or petroleum. As a result, it has been argued that biomass processing will occur at relatively small scales, drawing biomass from a distance as little as 15 miles around the plant. Th e corresponding rate of processing of biomass is equivalent to several hundred megawatts of power – at least an order of magnitude smaller than petro-leum refi neries.

In fact, there is an optimal size for biorefi neries since unit costs for processing go down while feedstock transportation costs go up as the plant size increases.1–3 Th e optimal size for diff erent kinds of biomass processing plants are not estab-lished but they are expected to depend upon the nature of biomass processed and the kind of processes employed.

Th is paper expands upon a previous study by the authors which compared the economics of advanced biorefi neries based on the biochemical and thermochemical platforms.4 While the previous paper compared capital and operating costs of comparably sized plants, the present paper explores the factors that infl uence the optimal size of biorefi neries and calculates the optimal size of several kinds of biofuels plants and the resulting unit costs for biofuels produced in them. Technologies examined include dry grind corn to ethanol,5 lignocellusic ethanol via enzymatic hydrolysis,6 gasifi cation and upgrading to hydrogen, methanol, and Fischer Tropsch liquids,7,8conversion of lignocellulosic biomass to mixed alcohols,9 and fast pyrolysis of lignocellu-losic biomass to bio-oil.10

Background

Biochemical conversion of sugar cane or grain crops (partic-ularly corn) to ethanol is commercially available and widely

practiced in Brazil and the USA, respectively. Grain ethanol plants are generally classifi ed as wet or dry milling based on the grain pretreatment. Wet milling plants have the advan-tage of lower production costs and higher effi ciencies, but the grain ethanol industry is currently dominated by dry milling plants due to the lower capital and labor costs of the latter11 and is thus the focus of this study.

Figure 1 illustrates the fi ve major steps of dry milling of corn to ethanol: grinding to expose the starch to enzymes; liquefaction by the action of heat and enzymes; saccharifi ca-tion by enzymatic activity; fermentation of sugars by yeast to produce ethanol; and distillation to neat ethanol. A by-product of this process is distillers dried grains (DDGS), a fi ber and protein-rich material that is used for livestock feed.

Technical and economic data for the dry grind ethanol process are taken from McAloon et al.5 Th ey determined that a plant capacity of 25 million gallons of ethanol per year had a production cost of $0.88 per gallon of ethanol (2000 basis year).

As shown in Fig. 2, biochemical conversion of cellulosic biomass into ethanol consists of four major steps: chemical and mechanical pre-treatments to release sugars from hemicellulose and to make cellulose accessible to enzymes; acid or enzymatic hydrolysis of cellulose into glucose; fermentation of the resulting hexose and pentose to ethanol, and distillation to yield neat ethanol. A by-product of this process is lignin, which can be employed as boiler fuel.

Technical and economic data for biochemical conversion of biomass to ethanol are from the analysis by Hamelinck et al.5 which assumes dilute acid pretreatment and enzy-matic hydrolysis. Th e analysis found that a plant of 50 million gallons per year capacity would produce ethanol at a cost of $1.51 per gallon of ethanol (2005 basis year).

Figure 1. Process fl ow diagram for dry grind of corn to ethanol.

© 2007 Society of Chemical Industry and John Wiley & Sons, Ltd | Biofuels, Bioprod. Bioref. 1:191–200 (2007); DOI: 10.1002/bbb 193

Review: Optimal sizes of different kinds of biorefineries M Wright, RC Brown

Th ermochemical conversion of biomass off ers a diversity of pathways to a number of biofuels. In general, these path-ways are based on either gasifi cation of biomass to gaseous products or fast pyrolysis to liquid products. In both cases these products represent intermediates in the manufacture of transportation fuels.

Gasifi cation of lignocellulosic biomass yields syngas, which consists mostly of hydrogen (H2) and carbon monoxide (CO).12 A variety of catalytic and even biocatalytic schemes have been developed to upgrade syngas into alcohols, ethers, esters, and hydrocarbons. In this study, four fuels from syngas are considered: hydrogen, methanol, mixed alcohols (with the purpose of maximizing ethanol synthesis), and Fischer Tropsch liquids. Th e optimal H2:CO molar ratio diff ers for each synthesis route. Methanol synthesis favors a hydrogen-to-carbon monoxide molar ratio of 3:1 while Fischer Tropsch diesel is optimized at a ratio of 2.15:1.13 Mixed alcohols optimized for ethanol production employs a syngas ratio of 0.6:1.9

As illustrated in Fig. 3, gasifi cation routes to biofuels have four major operations in common: comminution of the feedstock; gasifi cation; gas cleaning to remove tar, particu-

late matter, and inorganic contaminants; and water–gas shift reaction to enrich hydrogen with respect to carbon monoxide.14 If pure hydrogen is the desired product, at least two stages of water–gas shift are employed followed by gas purifi cation to remove carbon dioxide. Conversion of biomass to methanol, mixed alcohols, or Fischer Tropsch liquids adds catalytic synthesis steps aft er the water–gas shift unit operation. Production of mixed alcohols further adds a distillation step to separate fuel ethanol from the other alcohols. In all cases, unit operations are highly integrated to achieve heat recovery and utilize waste heat in electricity production.

Technical and economic data for hydrogen and methanol production are from Hamelinck and Faaij,7 who assumed biomass throughput equivalent to 400 MW thermal and pressurized, oxygen-blown gasifi cation. Th is represents a hydrogen plant yielding 182 million gallons per year at a cost of $0.24 per gallon of hydrogen or a methanol plant yielding 87 million gallons per year at a cost of $0.62 per gallon (2002 basis year). Technical and economic data for production of Fischer Tropsch liquids are from a study by Tijmensen et al.8 which assumes pressurized, oxygen-blown

Figure 2. Process fl ow diagram for biochemical conversion of lignocellulosic biomass to ethanol.

Figure 3. Process fl ow diagram for gasifi cation to hydrogen, methanol, mixed alcohols, or Fischer Tropsch liquids.

194 © 2007 Society of Chemical Industry and John Wiley & Sons, Ltd | Biofuels, Bioprod. Bioref. 1:191–200 (2007); DOI: 10.1002/bbb

M Wright, RC Brown Review: Optimal sizes of different kinds of biorefineries

gasifi cation. Th is analysis found that a plant of 35 million gallons per year capacity would produce Fischer Tropsch liquids at a cost of $2.37 per gallon (2002 basis year). Tech-nical and economic data for production of mixed alcohols are taken from a study by Phillips et al.9 which assumed atmospheric, indirectly heated gasifi cation. Th is analysis found that a plant capacity of 61.8 million gallons of ethanol per year would have a production cost of $1.01 per gallon (2005 basis year).

Fast pyrolysis is another thermochemical route to liquid fuels. Rapid heating of biomass at moderate temperatures (450–500°C) in the absence of oxygen directly yields a liquid product. As illustrated in Fig. 4, bio-oil production involves four major operations: comminution of the feedstock; fast pyrolysis; gas cleaning to remove particulate matter; and recovery of bio-oil. Although bio-oil can be directly used as boiler fuel or even fi red in certain kinds of engines, it is not suitable as a transportation fuel without further upgrading. However, the optimal size of a plant to produce bio-oil is included in this study because it illustrates the opportunities to explore distributed processing of biomass. Technical and economic data are taken from Ringer et al.10 Th eir analysis found that a plant of 28 million gallons per year capacity (assumed 90% capacity factor) would have a production cost of $0.34 per gallon of bio-oil (2003 basis year).

One option for upgrading bio-oil is hydrocracking it to diesel fuel and gasoline.15 Another option for upgrading bio-oil is gasifi cation followed by catalytic synthesis to Fischer Tropsch liquids. However, insuffi cient technical and economic data is currently available to include these biofuel options in the present study.

Methodology

Th e petroleum-based motor fuels industry is characterized by giant refi neries, processing petroleum at a rate equivalent to gigawatts of power. Th is situation exists because plant operating costs are driven by economies of scale, which

causes plant operating costs to increase more slowly as a plant gets bigger. Specifi cally, the operating cost (excluding fuel costs) of a plant, CP, scales with plant capacity, M, according to the power law:

CP � CPo (M/Mo)n (1)

where CPo is the plant operating cost for a plant of capacity Mo and n is a power law exponent less than unity, oft en assumed to be 0.6 (the ‘sixth-tenth’ rule). Ngyuen and Prince2 suggest that n is likely to be in the range 0.6 to 0.8 while Jenkins,3 in analyzing the optimal size of electric power plants, argues that this range is only appropriate for biomass power plants of size less than 50 MW electrical. He cites a study by Fisher et al.16 that supports values of n as large as 0.93–0.94 for coal-power plants in the size range 100–1400 MW.

Th e total cost, CT, for producing a quantity M of motor fuel from fossil fuel is the sum of the cost of plant operations, CP, the cost of feedstock at the mine mouth, CF, and the cost of feedstock delivery (transportation), CD:

CT � CP � CD � CF (2) � CPo (M/Mo)n � CFo (M/Mo) � CDo (M/Mo) (3)

where CFo is the feedstock cost and CDo is the feedstock delivery cost for a plant of capacity Mo. Th e unit cost for the resulting motor fuel ($ per gallon) is determined by dividing through by the plant capacity M:

(CT /M)� (CPo/Mon) Mn�1 � (CFo/Mo) � (CDo /Mo) (4)

Th us, for n less than unity, it is evident that the unit cost of the motor fuel from fossil fuel feedstocks decreases as the plant gets bigger, without limit.

Th e case is more complicated for biomass fuel since it does not come from a single mine mouth located a fi xed distance from the plant but is dispersed over a large area surrounding the plant. Th us, delivery cost for a unit of biomass fuel increases as the capacity of the plant increases because

Figure 4. Process fl ow diagram for production of bio-oil from fast pyrolysis.

© 2007 Society of Chemical Industry and John Wiley & Sons, Ltd | Biofuels, Bioprod. Bioref. 1:191–200 (2007); DOI: 10.1002/bbb 195

Review: Optimal sizes of different kinds of biorefineries M Wright, RC Brown

the biomass must be delivered from increasingly greater distances. Such an analysis has been explored by several researchers.1–3

Th e cost of biomass delivery is proportional to both trans-port distance and the quantity of biomass transported. Since the amount of biomass around a plant increases as the square of distance, D, from the plant, the cost of delivery is:

CD � CDo(D/Do)(M/Mo) � CDo(M/Mo)0.5(M/Mo) � CDo(M/Mo)1.5 � CDo(M/Mo)m (5)

where Do is the delivery distance for a plant of capacity Mo and m has been substituted for 1.5 to allow for some varia-tion in this power law. Nguyen and Prince2 argue that the power law exponent m might be as large as 2 if available land for biomass becomes increasingly sparse with distance from a plant, although in most cases 1.5 is probably a realistic value for m.

Th e total cost for producing a quantity of biomass-derived motor fuel is the cost of plant operations, CP, the cost of biomass fuel delivery, CD, and the cost of biomass fuel at the farm gate, CF:

CT � CP � CD � CF (6) � CPo (M/Mo)n � CDo(M/Mo)m � CFo(M/Mo) (7)

Th is expression is divided by Mo to obtain the unit cost for motor fuel produced from biomass:

(CT /M)� (CPo/Mon) Mn � 1 � (CDo/Mo

m) Mm � 1 � (CFo/Mo) (8)

Since n – 1 is less than zero and m – 1 is greater than zero, the fi rst term on the right-hand side of Eqn (8) decreases with plant capacity, while the second term increases with plant capacity. Th us, there is an optimum plant size to achieve minimum unit cost of motor fuel derived from biomass, which is obtained by diff erentiating Eqn (8), setting it equal to zero, and solving for Mopt/Mo. Th e result obtained is:

(Mopt/Mo) � {[(1 � n)/(m � 1)](CPo/CDo)}1/(m – n) (9)

It is evident from this expression that if operating costs CPo for a baseline plant of size Mo are much greater than the transportation cost CDo of biomass, then the optimum plant

size Mopt will be much greater than the size of the baseline plant.

Further insight into plant size optimization can be found by rearranging Eqn (9) as follows:

CDo(Mopt/Mo)m / CPo(Mopt/Mo)n � (1 – n)/(m – 1) (10)

Th e quantity on the left -hand side of Eqn (10) is recognized as the ratio of cost of delivery of biomass to the cost of plant operations, R, under the conditions of optimization; thus:

Ropt � (CD/CP)opt � (1 � n)/(m � 1) (11)

Th is indicates that, for a plant sized to give the minimum unit cost of motor fuel, the ratio of delivery cost to operating cost depends only on the power law exponents m and n. For example, assuming n is 0.6 (the sixth-tenth rule of economies of scale) and m is 1.5, delivery costs will equal 80% of plant operating costs in an optimally sized plant for minimum unit cost of motor fuel.

Once values of n and m are selected, the unit cost of biofuel as a function of plant of size M can be calculated from Eqn (8) and from knowledge of the cost of plant operations, the cost of feedstock, and the cost of feedstock delivery for a plant of arbitrary baseline size Mo. Th e cost of plant operations, CPo, for baseline cases of the various kinds of biofuel plants was obtained from the literature cited in the background section of this paper. All plant operating costs were infl ation adjusted to place them in the common basis year of 2005. Th e farm gate cost of feedstock was assumed to be $75.71 ton�1 ($2.12 bushel�1) for corn (the price in the basis year 2005) and $40 ton�1 of lignocellulosic biomass.

Th e cost of delivering feedstock from the farm gate to the plant gate is calculated from:

CDo = CDU . __

D .F (12)

where CDU is the unit cost for feedstock delivery (dollars per ton per mile),

__ D is the average delivery distance of feedstock,

and F is the tons of feedstock delivered annually to the plant. Th e unit cost of feedstock delivery for corn grain17 is $0.018 bushel�1 mile�1 and for lignocellulosic biomass18 is $0.71 ton�1 mile�1.

196 © 2007 Society of Chemical Industry and John Wiley & Sons, Ltd | Biofuels, Bioprod. Bioref. 1:191–200 (2007); DOI: 10.1002/bbb

M Wright, RC Brown Review: Optimal sizes of different kinds of biorefineries

For biomass uniformly distributed around a processing plant, the maximum radius rmax around the plant from which feedstock must be delivered is given by:

rmax � √ ___ F ___ π fY (13)

where f is the fraction of the acreage around a plant devoted to feedstock production (assumed to be 60% in this study) and Y is the annual yield of feedstock (assumed to 140 bushel acre�1 or 3.92 ton acre�1 for corn grain and 5 ton acre�1 for lignocellulosic biomass).

Th e average (land area-weighted) radius from which feed-stock is obtained around the plant is two-thirds of rmax. Th e actual distance traveled by a truck delivering biomass is greater than the radial (straight-line) distance to the plant depending upon the nature of the road network. To account for this additional distance, a ‘tortuosity factor’ τ is defi ned as the ratio of actual distance traveled to the straight-line distance from the plant. Tortuosity factors can be about 1.2 for developed agriculture regions where roads are laid out in rectangular grids or as great as 3.0 for less developed regions. Th e present analysis assumes the tortuosity factor is 1.5. Based on these adjustments, the average delivery distance is:

D � 2 _ 3 τ √ ___

F ___ π fY (14)

Table 1 details the average delivery distances and annual cost of delivering biomass to the baseline plants calcu-lated according to Eqns (14) and (12), respectively. Average delivery distance ranges between 5.5 miles and 11.8 miles for these seven base cases. Th e table also includes the volumetric (lower) heating values of the biofuels produced for each process because comparisons among the diff erent fuels are done on the basis of gallons gasoline equivalency (gge).

Table 2 is a compilation of plant size, capital cost (infl ation adjusted to 2005), plant operating costs (infl ation adjusted to 2005), unit cost of biofuels production, and R value of the various baseline biofuel plants to support the subsequent analysis of this study. Th e fact that none of the R values approaches the optimal value of 0.8 suggests that these base-line plants are far from the size that yields the minimum unit cost of biofuels production.

Results

Table 3 summarizes the results of the plant size optimiza-tion for processing cost power law exponent, n, and delivery cost power law exponent, m, equal to 0.6 and 1.5 respectively (corresponding to R equal to 0.8). Th e biofuels plants are listed in order of increasing optimal plant capacity in terms of gallons of gasoline equivalent (gge): corn grain ethanol, fast pyrolysis to bio-oil, biochemical production of

Table 1. Average delivery distances and delivery costs for baseline cellulosic biomass plants.

PlantBiofuel heating value (MJ L�1)

Plant size (million ggea)

Biomass input (million tons)

Average delivery distanceb (miles) CDo

c (million $)Grain ethanol 21 16.7 0.25 6.44 1.16

Biochemical cellulosic ethanol

21 33.5 0.66 11.8 5.51

Gasifi cation to hydrogen (liquid)

6 47.9 0.67 10.5 4.99

Gasifi cation to methanol 16 43.4 0.67 10.5 4.95

Gasifi cation to mixed alcohols

21 41.2 0.77 11.3 6.15

Gasifi cation to Fischer Tropsch diesel

36 40.4 0.64 10.3 4.66

Fast pyrolysis to bio-oil 19.6 17.3 0.18 5.46 0.70a Gallons gasoline equivalency.b Assumes feedstock yield of 3.92 ton acre�1 for corn grain and 5 ton acre�1 for lignocellulosic biomass; land utilization factor of 60%; and

tortuosity factor of 1.5.c Assumes feedstock delivery cost of $0.018 bushel�1 mile�1 for corn grain and $0.71 ton�1 mile�1 for lignocellulosic biomass.

© 2007 Society of Chemical Industry and John Wiley & Sons, Ltd | Biofuels, Bioprod. Bioref. 1:191–200 (2007); DOI: 10.1002/bbb 197

Review: Optimal sizes of different kinds of biorefineries M Wright, RC Brown

Table 3. Optimum plant annual capacity and operating costs (R � 0.8).

Process

M(millionggea)

Biomassinput

(milliontons)

Capitalcostb

(million $)

Operating costs (million $)CT/M ($

per ggea)

Volumetric annual plant

capacity(milliongallons)

CT/M($ per gallon)CP

b CD CF CT

Grain ethanol 79 1.18 70.9 14.9 12 90 117 1.48 119 0.98

Fast pyrolysis to bio-oil 104 1.08 88.4 12.7 10.2 43.1 66.0 0.63 168 0.39

Biochemicalcellulosicethanol 240 4.72 957 132 105 189 426 1.78 359 1.19

Gasifi cation to methanol 346 5.34 884 140 112 214 466 1.35 694 0.67

Gasifi cation to hydrogen 365 5.11 826 131 105 204 440 1.21 1387 0.32

Gasifi cation to mixed alcohols 418 7.81 551 249 199 312 760 1.82 627 1.21

Gasifi cation to Fischer Tropsch diesel 486 7.69 1516 243 194 308 745 1.53 422 1.77a Gallons gasoline equivalency.b Adjusted to 2005 basis year.

Table 2. Base case sizes and annual costs.

Process

M0(millionggea)

Biomassinput

(milliontons)

Capitalcostsb

(million $)

Operating costs (million $) CT/Mo($ per ggea) RcCPo

b CDo CFo CTo

Grain ethanol 16.7 0.25 27.9 5.86 1.16 18.9 25.9 1.55 0.20

Biochemicalcellulosic ethanol 33.5 0.66 294 40.5 5.51 26.4 72.4 2.16 0.14

Gasifi cation to hydrogen (liquid) 47.9 0.67 244 38.8 4.99 26.8 70.6 1.47 0.13

Gasifi cation to methanol 43.4 0.67 254 40.4 4.95 26.8 72.2 1.66 0.12

Gasifi cation to mixed alcohols 41.2 0.77 137 61.9 6.15 30.8 98.9 2.40 0.10

Gasifi cation to Fischer Tropsch diesel 40.4 0.64 341 54.6 4.66 25.6 84.8 2.10 0.09

Fast pyrolysis to bio-oil 17.3 0.18 30 4.35 0.695 7.20 12.2 0.71 0.16a Gallons gasoline equivalency.

b Adjusted to 2005 basis year.

c R is the ratio of cost of delivery of biomass to the cost of plant operations.

cellulosic ethanol, and gasifi cation of lignocellulosic biomass to methanol, hydrogen, mixed alcohols, and Fischer Tropsch diesel. Th e relatively low unit cost of plant operations rela-tive to feedstock delivery for corn grain ethanol allows it to

be built at an optimal scale of 79 million gge. In compar-ison, biochemical cellulosic ethanol requires a plant of 240 million gge capacity and the gasifi cation-based biofuel plants range in size from 346 million gge to 486 million gge

198 © 2007 Society of Chemical Industry and John Wiley & Sons, Ltd | Biofuels, Bioprod. Bioref. 1:191–200 (2007); DOI: 10.1002/bbb

M Wright, RC Brown Review: Optimal sizes of different kinds of biorefineries

to produce biofuel at minimum unit cost. On the basis of gallons of gasoline equivalency, optimally sized plants based on gasifi cation are 50–100% larger than biochemical cellu-losic ethanol plants. With the exception of the pyrolysis to bio-oil plant, all optimally-sized biofuel plants using ligno-cellulosic biomass as feedstock consume 4.7–7.8 million tons of biomass annually compared to 1.2 million tons of corn for the optimally sized grain ethanol plant.

Capital costs range from a low of $71 million for the 79 million gge grain ethanol plant to $1.52 billion for the 486 million gge Fischer-Tropsch diesel plant. All of the cellulose-to-biofuels plants exceed one-half billion dollars to build at their optimal size (which are larger than 240 million gge of biofuels). Th e unit cost of bio-oil is only $0.63 per gge for the optimally scaled plant although bio-oil is not suitable as motor fuel without additional upgrading. It is followed by hydrogen at $1.21, methanol at $1.35, grain ethanol at $1.48, Fischer Tropsch diesel at $1.53, biochemical cellu-losic ethanol at $1.78, and ethanol from the mixed alcohols process at $1.82. Among the cellulosic biofuel options, ethanol, whether produced biochemically or thermochemi-cally, is the most expensive to produce.

Th e total production costs for each of the plants (with the exception of the fast pyrolysis process) are plotted as

functions of plant size in Fig. 5. As expected, total costs at fi rst decline rapidly as plant capacity increases. Even-tually total costs reach a minimum value. Th ese minima are remarkably shallow. With the exception of the grain ethanol plant, capacity can be varied by 100 million gge or more around the optimal capacity without substantially aff ecting unit cost of biofuel production. Below about 50 million gge, or 75 million gallons ethanol, the grain ethanol process yields the lowest unit cost of biofuel. At larger plant capacities, grain ethanol becomes more expensive than hydrogen (at 50 million gge), methanol (at 100 million gge), and Fischer Tropsch diesel (at 325 million gge). Ethanol produced from lignocellulosic biomass, either biochemically or thermochemically is more expensive than ethanol from $2.12 bushel–1 corn grain even at plant capacities of one billion gge.

Relatively optimistic assumptions were made about lignocellulosic biomass yield (5 ton acre–1) and the percentage of land around the plant that would be devoted to growing this biomass (60%). Reductions in either yield or the percentage of land use could signifi cantly reduce the optimal size of biomass processing plants as calculated in this study. For example, substituting an agriculture residue such as corn stover for a dedicated energy crop might reduce

Figure 5. Unit cost of biofuels production versus plant size.

© 2007 Society of Chemical Industry and John Wiley & Sons, Ltd | Biofuels, Bioprod. Bioref. 1:191–200 (2007); DOI: 10.1002/bbb 199

Review: Optimal sizes of different kinds of biorefineries M Wright, RC Brown

the yield of lignocellulosic biomass to 2.5 ton acre–1. Alter-natively, the number of producers surrounding the plant willing to produce a dedicated energy crop like switchgrass might control only 30% of the land surrounding a processing plant. In either case, the optimal size of plants processing lignocellulosic biomass would be reduced by about 30%.

Another scenario that would result in smaller optimally sized processing plants is illustrated in Fig. 6, which is a plot of optimum plant size as a function of the power law exponent n that describes the scaling of feedstock processing costs with plant size. For low values of n, the optimal plant capacity increases with increasing n. Aft er a maximum optimum plant size is reached, optimal plant sizes decrease rapidly with increasing n indicating that small biofuels plants could be economically built if linear scaling (n approaching unity) prevailed. In principle, linear scaling could be approached by mass production of small-scale biofuels plants whereas today plants are custom designed and constructed. However, the authors are not aware of any such instances of mass production that demonstrate the practicality of this principle. Of course, any technology advances that reduce the capital cost and hence capital charges in the plant operating costs would also tend to reduce the optimal size of plants.

On the other hand, it is also easy to envision changes to the grain ethanol plant that would increase its optimal size and reduce the cost of grain ethanol production. For example, many ethanol producers are considering the addi-tion of front-end separation of corn oil to their dry-grind ethanol plants, which would increase plant processing costs. Transportation costs for these plants could be signifi cantly reduced if grain was shipped by rail or barge instead of by truck. According to Eq. 9, these factors will increase the optimal size of a grain ethanol plant to larger than the 79 million gge calculated in this study and lower the cost of grain ethanol.

Conclusions

Unlike fossil fuel processing plants, for which unit cost of fuel product decreases as the plant gets larger, optimal biomass processing plant capacities that achieve the lowest unit cost of biofuel production are expected. Th e optimal capacity depends upon the value of power law exponents used in scaling relationships that describe the costs of feed-stock processing and delivery, respectively, and upon the relative cost of feedstock processing compared to feedstock delivery.

Figure 6. Effect of processing cost scale factor n on the optimal size of biofuels plants.

200 © 2007 Society of Chemical Industry and John Wiley & Sons, Ltd | Biofuels, Bioprod. Bioref. 1:191–200 (2007); DOI: 10.1002/bbb

M Wright, RC Brown Review: Optimal sizes of different kinds of biorefineries

Data from the published literature were used to evaluate the optimal size of several biorefi nery concepts. In the order of their increasing optimal plant size these are: corn grain ethanol, fast pyrolysis to bio-oil, biochemical production of cellulosic ethanol, gasifi cation to methanol, gasifi cation to hydrogen, gasifi cation to mixed alcohols, and gasifi ca-tion to Fischer Tropsch diesel. Optimally sized plants based on gasifi cation are 50–100% larger (gge basis) than the biochemical cellulosic ethanol plant. Biorefi neries converting lignocellulosic biomass into transportation fuels are opti-mally sized in the range 240 to 486 million gge per year and require 4.7–7.8 million tons of biomass annually compared to 1.2 million tons of corn grain for an optimally sized grain ethanol plant. Capital costs for advanced biofuel plants are similarly much greater than for a grain ethanol plant, ranging from $551 million for a mixed alcohols plant to $1.52 billion for a Fischer Tropsch diesel plant compared to $70.9 million for the grain ethanol plant. Among the cellulosic biofuel options, ethanol, whether produced biochemically or thermochemically, is the most expensive to produce.

Th e minima in the unit cost of cellulosic biofuel versus plant capacity are so shallow that, with the exception of the fast pyrolysis plant, plant size can be varied by as much as 100 million gge around the optimal capacity without substantially aff ecting unit cost of biofuels production.

Although optimally sized biorefi neries based on lignocel-lulosic biomass are predicted in this study to process four to seven times as much feedstock as grain ethanol plants, three factors could reduce the optimal size of these advanced biorefi neries. First, reductions in either biomass yield or the percentage of land surrounding a plant dedicated to biomass crops would signifi cantly increase the distance required to collect biomass, making it too expensive to build as large a plant as projected in this study. Second, making the power law for biomass processing nearly linear through mass production of small-scale biofuels plants might reduce the optimal size of these plants. Finally, technology advances that reduce capital costs of the plant would reduce the size of plant that produces biofuel at the lowest unit cost.

Acknowledgments

Th is study was supported by the Iowa Energy Center under Contract No. 0606 and the USDA under Contract No. 5819356654.

References 1. Overend RP, The average haul distance and transportation work

factors for biomass delivered to a central plant. Biomass 2:75–79

(1982).

2. Nguyen MH and Prince RGH, Simple rule for bioenergy conversion plant

size optimisation: bioethanol from sugar cane and sweet sorghum.

Biomass Bioenergy 10 (5-6):361–365 (1996).

3. Jenkins BM, A Comment on the optimal sizing of a biomass utilization

facility under constant and variable cost scaling. Biomass Bioenergy

13:1–9 (1997).

4. Wright M and Brown R, Comparative economics of biorefi neries based

on the biochemical and thermochemical platforms, Biofuels Bioprod

Bioref 1:49–56 (2007).

5. McAloon A. Taylor F, Yee W, Ibsen K and Wooley R, Determining

the cost of producing ethanol from corn starch and lignocellulosic

feedstocks. National Renewable Energy Laboratory, Report, October

(2000).

6. Hamelinck CN, van Hooijdonk G and Faaij APC, Ethanol from

lignocellulosic biomass: techno-economic performance in

short-, middle-, and long-term. Biomass Bioenergy 28:384 (2005).

7. Hamelinck CN and Faaij A, Future prospects for production of methanol

and hydrogen from biomass. J Power Sources 111:1 (2002).

8. Tijmensen JAM, Faaij APC and Hamelinck CN, Exploration of the

possibilities for production of Fischer Tropsch liquids and power via

biomass gasifi cation. Biomass Bioenergy 23:129 (2002).

9. Phillips S, Aden A, Jechura J, Dayton D and Eggeman T, Thermoche-

mical ethanol via indirect gasifi cation and mixed alcohol synthesis of

lignocellulosic biomass. NREL Report NREL/TP-510-41168 April

(2007).

10. Ringer M, Putsche V and Scahill J, Large-scale pyrolysis oil production:

a technology assessment and economic analysis. NREL Report NREL/

TP-510-37779 November (2006).

11. Brown RC, Biorenewable Resources: Engineering New Products from

Agriculture. Blackwell Publishing: Ames, IA (2003)

12. Reed TB, Biomass Gasifi cation: Principles and Technology. Noyes Data

Corporation/Noyes Publications: January (1981).

13. Spath PL, Dayton DC, Preliminary screening – technocal and economic

assessment of synthesis gas to fuels and chemicals with emphasis on

the potential for biomass-derived syngas. NREL Report NREL/TP-

510-34929 December (2003).

14. Probstein RF and Hicks RE, Synthetic Fuels. Dover Publications, Inc.:

Mineola, NY (2006).

15. Marinangelli R, Marker T, Petri J, Kaines T, McCall M, Mackowiak et al.

Opportunities for biorenewables in oil refi neries. Department of Energy

Final Technical Report (2005).

16. Fisher CF, Paik S Jr and Schriver R, Power plant economy of scale and

cost trends – further analyses and review of empirical studies. ORNL/

Sub-85-7685/1-11, DE86-014410, Oak Ridge National Laboratory, Oak

Ridge, TN (1986).

17. Edwards W and Smith D, Iowa Farm Custom Rate Survey, Ag Decision

Maker, File A3-10 (2007).

18. Birrell S, Transportation cost for herbaceous biomass. Personal

communication, Iowa State University, July (2007).